mattm@chromium.org src/net/cert/OWNERS Information about the certificate algorithms and sizes in use on the web, to examine compliance with the CA/Browser Forum requirements and security best practice. This histogram considers the notBefore as the issuance date, for purposes of what requirements apply. {BRValidity} Records the size in bits of the key for certificates with chain position {ChainPosition} and algorithm {Algorithm}. seblalancette@chromium.org nsatragno@chromium.org src/crypto/OWNERS Collected on Mac and iOS when a {Operation} operation failed for a SecureEnclave-protected ECDSA key. Represents the OSStatus code returned by the API corresponding with that operation. seblalancette@chromium.org nsatragno@chromium.org src/crypto/OWNERS Collected on Mac when a {Operation} operation failed for a SecureEnclave-protected ECDSA key. Represents the OSStatus code returned by the API corresponding with that operation. Warning: this histogram was expired from 2025-10-06 to 2025-12-08; data may be missing. Note: This histogram should be removed upon expiration and replaced by Crypto.SecureEnclaveOperation.Apple.{Operation}.Error kristianm@chromium.org agl@chromium.org Collected on Windows at startup. Records status for various virtual TPM operations that Chrome artificially triggers on startup for measurement purposes. Note that if new key creation fails, wrapped key creation and message signing will not take place. kristianm@chromium.org wfh@chromium.org Collected on Windows at startup. Records status for various TPM operations that Chrome artificially triggers on startup for measurement purposes. Note that if new key creation fails, wrapped key creation and message signing will not take place. kristianm@chromium.org agl@chromium.org Error code returned by NCryptCreatePersistedKey. One of the error codes from https://learn.microsoft.com/en-us/windows/win32/com/com-error-codes-4. kristianm@chromium.org agl@chromium.org Error code returned by NCryptFinalizeKey. One of the error codes from https://learn.microsoft.com/en-us/windows/win32/com/com-error-codes-4. kristianm@chromium.org agl@chromium.org Error code returned by NCryptOpenKey. One of the error codes from https://learn.microsoft.com/en-us/windows/win32/com/com-error-codes-4. kristianm@chromium.org agl@chromium.org Error code returned by NCryptOpenStorageProvider. One of the error codes from https://learn.microsoft.com/en-us/windows/win32/com/com-error-codes-4. kristianm@chromium.org agl@chromium.org Collected on Windows at startup. Records status for various virtual TPM actions that Chrome artificially triggers on startup for measurement purposes. Note that if new key creation fails, wrapped key creation and message signing will not take place. jdoerrie@chromium.org src/crypto/OWNERS Collected on Windows when tbs.dll fails to load during TPM certification. Records the HRESULT from loading the library. jdoerrie@chromium.org src/crypto/OWNERS Collected on Windows when Tbsip_Submit_Command fails during TPM certification. Represents the TBS_RESULT returned by the API. jdoerrie@chromium.org src/crypto/OWNERS Collected on Windows during TPM2_Certify verification. Records the TPM_ALG_ID of the hash algorithm unconditionally for all operations. jdoerrie@chromium.org src/crypto/OWNERS Collected on Windows to record the verification result of a TPM2_Certify response. Records whether the signature was valid or the specific error. jdoerrie@chromium.org src/crypto/OWNERS Collected on Windows during TPM2_Certify verification. Records the TPM_ALG_ID of the signature algorithm unconditionally for all operations. jdoerrie@chromium.org src/crypto/OWNERS Collected on Windows to record the result of property extraction before submitting a {Command} command. Represents the SECURITY_STATUS returned by the API. jdoerrie@chromium.org src/crypto/OWNERS Collected on Windows to record the result of parsing the response of a {Command} command. Represents the ParseResult, which is Ok on success or the specific ParseError that occurred. jdoerrie@chromium.org src/crypto/OWNERS Collected on Windows to record the result of a {Command} command at the TPM level. Represents the TPM 2.0 response code returned by the TPM. seblalancette@chromium.org kristianm@chromium.org src/crypto/OWNERS Collected on Windows when a {Operation} operation failed {SignatureAlgorithm}. Represents the Win32 status returned by the API corresponding with that operation. kristianm@chromium.org wfh@chromium.org Collected on Windows at startup. Records status for various TPM actions that Chrome artificially triggers on startup for measurement purposes. Note that if new key creation fails, wrapped key creation and message signing will not take place. agl@chromium.org davidben@chromium.org wfh@chromium.org Collected on Windows at startup. Records whether a TPM is present and, if so, what the best supported algorithm is. kristianm@chromium.org agl@chromium.org Collected on Windows at startup. Records whether a virtual and/or hardware TPM is present. alexilin@chromium.org drubery@chromium.org chrome-counter-abuse-alerts@google.com Records the time that passed between starting a TPM operation on an unexportable_keys::UnexportableKeyService and completing the operation (either successfully or not). This time includes both the time spent waiting in a task queue and the execution time. These sub-intervals are covered by Crypto.UnexportableKeys.BackgroundTaskQueueWaitDuration and Crypto.UnexportableKeys.BackgroundTaskRunDuration histograms correspondingly. Includes operations that were scheduled with {Priority}. Recorded right after a TPM operation completes. alexilin@chromium.org drubery@chromium.org chrome-counter-abuse-alerts@google.com Records the time spent waiting in a task queue for the subset of TPM operations that are asynchronously scheduled through unexportable_keys::UnexportableKeyService. This duration is a sub-interval of Crypto.UnexportableKeys.BackgroundTaskDuration. Includes operations that were scheduled with {Priority}. Recorded right before a TPM operation is sent to the background thread for execution. alexilin@chromium.org drubery@chromium.org chrome-counter-abuse-alerts@google.com Records the result of the subset of TPM operations that are scheduled asynchronously through unexportable_keys::UnexportableKeyService. In some error cases, the result might be returned synchronously. Recorded right after a TPM operation completes. Recorded only for {TaskType} originating from {TaskOrigin}. alexilin@chromium.org drubery@chromium.org chrome-counter-abuse-alerts@google.com Records the number retries it took to complete a TPM operation scheduled through unexportable_keys::UnexportableKeyService, only for operations that are completed with {Success}. Recorded right after a TPM operation completes. Recorded only for {TaskType}. alexilin@chromium.org drubery@chromium.org chrome-counter-abuse-alerts@google.com Records the time that passed between starting executing a TPM operation on a background thread and completing the operation (either successfully or not). Includes only a subset of TPM operations that are scheduled asynchronously through unexportable_keys::UnexportableKeyService. This duration is a sub-interval of Crypto.UnexportableKeys.BackgroundTaskDuration. Unlike Crypto.UnexportableKeys.BackgroundTaskDuration, this histogram is split by a task type instead of a task priority because the priority doesn't matter once a task starts executing. Only one task can be executed at the time due to TPM limitations. Includes {TaskType}. Recorded right after a TPM operation completes. alexilin@chromium.org drubery@chromium.org chrome-counter-abuse-alerts@google.com Records {MetricName} at the time of garbage collection triggered for {KeyDeletionClient}. Typically, the unexportable keys garbage collection runs with some delay after a profile startup but it also might be triggered in response to UnexportableKeyService deletion events. For example, a garbage collection will run when a profile is deleted. amraboelkher@chromium.org jdoerrie@chromium.org Recorded when a background task finishes generating a spare {KeyType} key to replenish the pool, regardless of success or failure. Records 'Success' (0, corresponding to kNoServiceErrorForMetrics in C++) or the specific ServiceError code on failure. amraboelkher@chromium.org jdoerrie@chromium.org Recorded when an asynchronous {KeyType} key generation request is received, immediately before checking the pool for a hit or triggering replenishment. Records the total {KeyType} key pool size. amraboelkher@chromium.org jdoerrie@chromium.org Recorded when a background task successfully generates a spare {KeyType} key to replenish the pool. Records the duration (in milliseconds) of the background task from its launch until completion. amraboelkher@chromium.org jdoerrie@chromium.org Recorded when an asynchronous {KeyType} key generation request completes (its callback is executed), regardless of whether the spare pool feature is enabled or disabled. Records the duration (in milliseconds) from the initial request until callback execution, covering both instantaneous pool hits and slow fallback background generations. amraboelkher@chromium.org jdoerrie@chromium.org Recorded when a consumer requests a {KeyType} key asynchronously. Records the outcome (hit, various miss reasons, or hardware support failure) of attempting to retrieve a pre-generated spare {KeyType} key from the background pool. kristianm@chromium.org agl@chromium.org Collected on Windows at startup. Records whether a virtual TPM is present and, if so, what the best supported algorithm is. morlovich@chromium.org olivierli@chromium.org For every http cache transaction with a pattern (see HttpCache.Pattern), the time from the first access to the disk cache until the transaction was done. {HttpCacheAccessToDoneCases} bashi@chromium.org net-dev@chromium.org The time taken an HttpCache::Transaction to be added to ActiveEntry. It roughly captures time to acquire a write lock for an disk cache. Recorded every time a transaction (re)starts. Not recorded if the cache is memory-backed. morlovich@chromium.org yhirano@chromium.org For http cache transactions in which a network request was sent, the time elapsed between beginning the request and sending it over the network; this is the time spent accessing the disk cache. {HttpCacheBeforeSendCases} yoichio@chromium.org chrome-loading@google.com Logged wether or not CreateBackend is called when HttpCache is created. This is valid when HttpCacheInitializeDiskCacheBackendEarly is enabled. yhirano@chromium.org net-dev@chromium.org The time taken to create a disk cache entry. Recorded when the entry is created for the first time in the HttpCacheTransaction. Not recorded if the cache is memory-backed. leimy@chromium.org chrome-owp-storage@google.com Records the outcome of considering whether to reject a cache entry based on in-memory hints. This is used to avoid accessing disk for entries that are likely to be unusable. Recorded every time we try to open a cache entry, but there is no active one{Type} anthonyvd@chromium.org chrome-catan@google.com Records the size of an HTTP cache entry's content section, every time the content of the entry is read from the cache. anthonyvd@chromium.org chrome-catan@google.com Records the size of an HTTP cache entry's response info section, every time the content of the entry is read from the cache. anthonyvd@chromium.org chrome-catan@google.com Records the total size (header + content) of an HTTP cache entry, every time the content of the entry is read from the cache. horo@chromium.org blink-network-stack@google.com Whether the HttpCache was forced to reset upon creation, effectively clearing out its previous content. This is recorded when the cache is created for the browsing session. Warning: this histogram was expired from June 2021 to June 2026; data may be missing. bashi@chromium.org blink-network-stack@google.com Whether or not a response has `Cache-Control: no-store`. Recorded when a HTTP transaction is finished. This histogram is for {Type}. shivanisha@chromium.org jkarlin@chromium.org yoichio@chromium.org chrome-loading@google.com Max file size allowed to be cached in the http cache. This is logged at the time the backend is created. Warning: this histogram was expired from 2022-02-20 to 2026-04-07; data may be missing. yoichio@chromium.org chrome-loading@google.com The time it takes to instantiate and synchronously execute the Load method of NoVarySearchCacheStorage::Loader on a worker pool thread. This is logged at the time the NVS backend is created. ricea@chromium.org net-dev@chromium.org The filesystem error from a failed attempt to delete the "snapshot.baj-new" temporary file from the {Directory} directory by net::NoVarySearchCacheStorage. This is only logged if the file exists and deleting it fails. The check is only done when creating an on-the-record NetworkContext. These errors should be rare and harmless. ricea@chromium.org net-dev@chromium.org The filesystem error from a failed attempt to create the Cache/No_Vary_Search directory by net::NoVarySearchCacheStorage. This is logged any time creating the directory fails. The directory is only created when opening a profile when the directory doesn't already exist. ricea@chromium.org net-dev@chromium.org The result from trying to ensure the existence of the Cache/No_Vary_Search directory by net::NoVarySearchCacheStorage. This is logged when an on-the-record NetworkContext is created. Anything other than "Create failed" is good. ricea@chromium.org net-dev@chromium.org Number of entries that were added to the NoVarySearchCache while the previously persisted cache was being loaded from disk. This is logged once per successful call to NoVarySearchCacheStorage::Load(), which is once per on-the-record NetworkContext. This can be considered an approximation of how many requests missed the benefit of the NoVarySearchCache due to the load being slow. ricea@chromium.org net-dev@chromium.org Number of entries that were loaded from a previously persisted NoVarySearchCache by a successful call to NoVarySearchCacheStorage::Start(). Logged once per successful call, which is once per on-the-record NetworkContext. ricea@chromium.org net-dev@chromium.org The result of parsing the No-Vary-Search value in the HTTP response header. This metric is emitted every time an HTTP response is written to the HTTP disk cache. ricea@chromium.org net-dev@chromium.org The filesystem error from a failed attempt to delete the {File} file from the {Directory} directory by net::NoVarySearchCacheStorage. This is logged when the {File} file is found to exist in both the old and new locations, and deleting it from the old location fails. This can only happen when opening the cache directory on creation of an on-the-record NetworkContext. Unless the quantity is very large these errors are harmless. ricea@chromium.org net-dev@chromium.org The filesystem error from a failed attempt to rename the {File} file from the {Directory} directory to the new location by net::NoVarySearchCacheStorage. This is logged when the {File} file is only found in the old location, and moving it from the old location fails. This can only happen when opening the cache directory on creation of an on-the-record NetworkContext. Unless the quantity is very large these errors are harmless. ricea@chromium.org net-dev@chromium.org A true value is logged whenever appending to the "journal.baj" file fails. Appending is done for every entry inserted into or erased from the cache. After an error, persistence is abandoned, and so this will be logged at most once per NetworkContext created. ricea@chromium.org net-dev@chromium.org The filesystem error from a failed attempt to create the "journal.baj" file by net::NoVarySearchCacheStorage. This is logged any time creating "journal.baj" fails. "journal.baj" is created when a new on-the-record profile is created, and recreated when it gets too full or when a bulk erase is done on the cache. After an error, persistence is abandoned, and so this will be logged at most once per NetworkContext created. ricea@chromium.org net-dev@chromium.org The filesystem error from a failed attempt to open the "journal.baj" file for reading by net::NoVarySearchCacheStorage. This is logged when loading "journal.baj" fails. "journal.baj" is loaded when an on-the-record profile is initialized, typically during browser startup. ricea@chromium.org net-dev@chromium.org A true value is logged whenever writing the magic number to a new "journal.baj" file fails. This is done whenever the file is created. After an error, persistence is abandoned, and so this will be logged at most once per NetworkContext created. ricea@chromium.org net-dev@chromium.org The filesystem error from a failed attempt to delete the legacy "no-vary-search" directory within "Cache/Cache_Data". This is logged when the deletion fails, which can only happen during initialization of an on-the-record NetworkContext. Unless the volume is very large these errors are harmless. ricea@chromium.org net-dev@chromium.org The result of attempting to load a previously persisted NoVarySearchCache from disk and start a new one. This is logged asynchronously after net::NoVarySearchCacheStorage::Start() is called, once per on-the-record NetworkContext creation. ricea@chromium.org net-dev@chromium.org The time between calling net::NoVarySearchCacheStorage::Start() and discovering that persistence has failed to be initialized, including cross-thread task posting overhead. See HttpCache.NoVarySearch.LoadResult for the cause of failure. Logged once per unsuccessful call to net::NoVarySearchCacheStorage::Start(). ricea@chromium.org net-dev@chromium.org The time between calling net::NoVarySearchCacheStorage::Start() and starting to use the previously-persisted data, including cross-thread task posting overhead. This is logged in any case where persistence is able to continue, even when data may have been lost. See HttpCache.NoVarySearch.LoadResult for the actual success rate in loading persisted data. Logged once per successful call to net::NoVarySearchCacheStorage::Start(). ricea@chromium.org net-dev@chromium.org The time spent executing net::NoVarySearchCache::Lookup(). This is recorded once for every call to the function {HitOrMiss}. {MainFrameOrNot} {BaseUrlMatchedOrNot} This will be used to evaluate the need for optimization. The histogram is not reported on Windows machines without high resolution clocks. ricea@chromium.org net-dev@chromium.org The time lost due to a successful lookup in the NoVarySearchCache that was later found to not be usable by Simple Cache in-memory hints. This happens when an entry is stale and cannot be revalidated, and the load flags do not permit the use of stale entries. In these cases the transaction is restarted with the original URL without using the NoVarySearchCache, and so time has been lost compared to if we'd never used the NoVarySearchCache at all. This histogram measures that lost time. This doesn't include the actual lookup time, but that is recorded in HttpCache.NoVarySearch.LookupTime.Hit and is usually an order of magnitude smaller. Recorded once for every cache transaction which is restarted due to getting a hit in the NoVarySearchCache that was then found to be unusable due to Simple Cache in-memory hints. This corresponds to the NoVarySearchUseResult case "Not suitable". This is a subset of the HttpCache.NoVarySearch.NotUsableLostTime2 histogram that is expected to be fast due to not including any asynchronous operations. The histogram is not reported on Windows machines without high resolution clocks. ricea@chromium.org net-dev@chromium.org The time lost due to a successful lookup in the NoVarySearchCache that was later found to not be usable, usually due to the disk entry either being missing or not conditionalizable. In these cases the transaction is restarted with the original URL without using the NoVarySearchCache, and so time has been lost compared to if we'd never used the NoVarySearchCache at all. This histogram measures that lost time. This doesn't include the actual lookup time, but that is recorded in HttpCache.NoVarySearch.LookupTime.Hit and is usually an order of magnitude smaller. Recorded once for {HostAndFrameType} cache transaction which is restarted due to getting a hit in the NoVarySearchCache that was later found to be unusable. This corresponds to the NoVarySearchUseResult cases "Not openable", "Incomplete body" and "Couldn't conditionalize". This histogram replaced HttpCache.NoVarySearch.NotUsableLostTime, which was logged too eagerly in some edge cases. ricea@chromium.org net-dev@chromium.org The result from checking for and possibly moving {File} from the {Directory} directory to a new location. This is logged when an on-the-record NetworkContext is created. "Source did not exist" and "Renamed" are good outcomes. "Rename failed" means the user had to start again from an empty NoVarySearchCache, but is otherwise harmless. See HttpCache.NoVarySearch.InitRenameError.* for the actual filesystem error. "Source deleted" probably means rename failed on a previous run but is harmless. "Deletion failed" means we left junk on the user's disk but is otherwise harmless. See HttpCache.NoVarySearch.InitDeleteError.* for the actual filesystem error. ricea@chromium.org net-dev@chromium.org The error code returned by base::ReplaceFile() inside the concrete implementation of NoVarySearchCacheStorageFileOperations::AtomicSave(). On Windows base::ReplaceFile() may be retried on failure, and if it is subsequently successful then this histogram will not be logged. See "HttpCache.NoVarySearch.ReplaceFileTryCount". Logged once per failure to replace the file in AtomicSave(). Not logged on success. ricea@chromium.org net-dev@chromium.org The number of calls to ReplaceFile() in ReplaceFileWithRetries() needed until one is successful. This should usually be 1, but may be more if there was interference from an anti-virus or other third-party program. Only logged on Windows because retries are only necessary on Windows. Not logged when ReplaceFile fails more than the maximum numer of retries. In that case "HttpCache.NoVarySearch.ReplaceFileError" will be logged with a value of IN_USE instead. Logged once for each successful call to AtomicSave() in the concrete implementation of NoVarySearchCacheStorageFileOperations on Windows. ricea@chromium.org net-dev@chromium.org Number of entries successfully read from the "journal.baj" file. This is logged once for every journal file that is successfully opened for reading and found to have a correct magic number. Journal load is triggered asynchronously by a call to net::NoVarySearchCacheStorage::Start() and happens once per on-the-record NetworkContext. ricea@chromium.org net-dev@chromium.org The filesystem error from a failed attempt to open the "snapshot.baf" file for reading by net::NoVarySearchCacheStorage. This is logged when loading "snapshot.baf" fails. "snapshot.baf" is loaded when an on-the-record profile is initialized, typically during browser startup. ricea@chromium.org net-dev@chromium.org The filesystem error from a failed attempt to write the "snapshot.baf" file by net::NoVarySearchCacheStorage. This is logged any time writing "snapshot.baf" fails. "snapshot.baf" is written when a new on-the-record profile is created, and rewritten when "journal.baj" gets too full or when a bulk erase is done on the cache. After an error, persistence is abandoned, and so this will be logged at most once per NetworkContext created. ricea@chromium.org net-dev@chromium.org The results of applying the No-Very-Search algorithm in the HTTP disk cache{HostAndFrame}. This metric is emitted once for every cache transaction when the transaction is complete. "Used" and "Validated" are the happy cases where applying No-Vary-Search reduced network traffic. "Not applied", "No match" and "URL Unchanged" it didn't help, but at minimal cost. In the case of "Updated", a different entry was written than would have been otherwise but nothing user-visible changed. "Not suitable" means it didn't help at a slightly higher cost. The other values mean it didn't help, and there was a cost of additional disk reads. This histogram replaced "HttpCache.NoVarySearch.UseResult", which was accidentally missing cache transactions that ended with a "Not Covered" reason. yhirano@chromium.org net-dev@chromium.org The time taken to open a disk cache entry. Recorded when the entry is created for the first time in the HttpCacheTransaction. Not recorded if the cache is memory-backed. ricea@chromium.org src/net/OWNERS For each http cache transaction that reported "Not covered" to the HttpCache.Pattern histogram, the reason why it was not covered. Logged once for every cache transaction that records a "Not covered" reason, in the destructor. {CachedResourceType} morlovich@chromium.org jkarlin@chromium.org For each http cache transaction, the recorded pattern. {CachedResourceType} hayato@chromium.org blink-network-stack@google.com Whether the freshness of a response is zero or not. Recorded for each http cache transaction. Emitted when a response is written into the HTTP cache. ashleynewson@chromium.org src/android_webview/OWNERS For each call to HttpCache::SetMaxBytes, records whether the disk cache backend was already created or is at least being created. This is emitted regardless of whether force_initialization is true or false for the SetMaxBytes call. HttpCache::SetMaxBytes is usually called as a result of using a setter in the Android WebView HTTP Cache Quota API. haihan@google.com valadkevich@google.com The time taken to initialize the disk cache at disk_cache::CreateCacheBackend level. bashi@chromium.org blink-network-stack@google.com Deprecated. Replaced by HttpCache.TotalDiskCacheTimePerTransaction.{AccessType}2. bashi@chromium.org blink-network-stack@google.com For each HTTP cache transaction, record total time spent {AccessType} the disk cache. Recorded when a transaction finished. This histogram does not record partial requests. yyanagisawa@chromium.org chrome-loading@google.com To optimize performance, the network service can handle `Accept-CH` frames directly without consulting the browser process (an optimization called 'offloading'). This histogram is recorded when that optimization fails. It identifies the specific client hint from the `Accept-CH` frame that was not in the network service's cache of hints the browser had previously identified as either allowed or disallowed for the origin. A value is recorded each time this mismatch occurs, forcing a fallback to the browser process. yyanagisawa@chromium.org chrome-loading@google.com The reason why AcceptCHFrameInterceptor::NeedsObserverCheck returns true. Recorded every time NeedsObserverCheck is called. It is recorded just before returning from the function. yyanagisawa@chromium.org chrome-loading@google.com When handling an `Accept-CH` frame, the network service can sometimes avoid a costly round-trip to the browser process by checking for required client hints itself (an optimization called 'offloading'). This histogram is recorded for each client hint that was successfully offloaded specifically because of a mechanism enabled by the AcceptCHFrameOffloadNotAllowedHints feature parameter. This mechanism allows offloading for hints that are persisted for an origin but currently disabled by Permissions Policy. This metric helps measure the effectiveness of this opt-in optimization. dschinazi@chromium.org src/net/OWNERS Breakdown of the locations when SetBrokenAlternateProtocol is called. dschinazi@chromium.org src/net/OWNERS Breakdown of how requests which could potentially make use of an alternate protocol use or don't use the protocol. dschinazi@chromium.org src/net/OWNERS Breakdown of how requests which could potentially make use of an alternate protocol use or don't use the protocol, for connections to Google servers. tbansal@chromium.org src/net/OWNERS Positive net error codes that failed alternative protocol requests end with. Recorded only when an alternative job fails, and the main job succeeds. horo@chromium.org src/net/OWNERS Positive net error codes that failed alternative protocol requests end with. Recorded only when an DNS alpn H3 job fails, and the main job succeeds. bashi@chromium.org src/net/OWNERS Information about the AlternativeService used for each request. Warning: this histogram was expired from 2024-04-03 to 2024-05-31; data may be missing. lukedeng@google.com dmurph@chromium.org src/net/OWNERS The outcome of attempting to register the Android default network callback in NetworkChangeNotifierAutoDetect. Recorded on Android 28+ (Pie+) on every registration attempt: - On application startup or when returning to the foreground (visible activity created/resumed). - When an activity regains window focus and triggers a self-healing retry because previous callback registration had failed. If an initial registration throws a SecurityException (e.g. transient UID permission delay on waking from Doze mode), a single immediate retry is attempted. This histogram records the final result (initial success with or without immediate retry, focus self-heal success with or without immediate retry, or failure categorized by exception type). lukedeng@google.com dmurph@chromium.org src/net/OWNERS The outcome of attempting to register the Android network request callback in NetworkChangeNotifierAutoDetect. Recorded on Android 21+ (Lollipop+) on every registration attempt: - On application startup or when returning to the foreground (visible activity created/resumed). - When an activity regains window focus and triggers a self-healing retry because previous callback registration had failed. If an initial registration throws a SecurityException (e.g. transient UID permission delay on waking from Doze mode), a single immediate retry is attempted. This histogram records the final result (initial success with or without immediate retry, focus self-heal success with or without immediate retry, or failure categorized by exception type). lukedeng@google.com dmurph@chromium.org src/net/OWNERS Categorization of exceptions caught during Android ConnectivityManager callback registration in NetworkChangeNotifierAutoDetect. Recorded once on each failed registration attempt when all retries have been exhausted. Categorizes the final exception into SecurityException, TooManyRequestsException, IllegalArgumentException, IllegalStateException, or other RuntimeException. lukedeng@google.com dmurph@chromium.org src/net/OWNERS The result of attempting to self-heal default network callback registration in NetworkChangeNotifierAutoDetect. Recorded when an Android activity regains window focus (onWindowFocusChanged with hasFocus=true) and the default network callback was previously in a failed state (e.g. due to Doze mode restrictions). Records whether the recovery attempt succeeded (RECOVERED) or failed (FAILED). lukedeng@google.com dmurph@chromium.org src/net/OWNERS The result of attempting to self-heal network request callback registration in NetworkChangeNotifierAutoDetect. Recorded when an Android activity regains window focus (onWindowFocusChanged with hasFocus=true) and the network request callback was previously in a failed state (e.g. due to Doze mode restrictions). Records whether the recovery attempt succeeded (RECOVERED) or failed (FAILED). leimy@chromium.org chrome-owp-storage@google.com Net error code from a blob fetch request. It's recorded as the abolute value of the raw error code when a blob fetch request fails, or as 0 when the request receives response. nidhijaju@chromium.org chrome-network-stack@google.com Records if a certificate verification result was provided from the cache or not for connections to {HostType}. Recorded after checking the certificate verification cache when CachingCertVerifier::Verify() is called. nidhijaju@chromium.org chrome-network-stack@google.com Records the time taken to {VerificationCompletionType} verify a certificate if not found in the certificate verification cache for connections to {HostType}. Recorded after a certificate is verified in the CachingCertVerifier. Note: This is similar to Net.CertVerifier_Job_Latency which is recorded in CoalescingCertVerifier. CachingCertVerifier wraps the CoalescingCertVerifier, so the main difference is if there are multiple verifications for the same params in a short time span, CoalescingCertVerifier records a single histogram event from the start of the first verification request, whereas this histogram includes the effect of averaging out the times if multiple requests are coalesced together. mattm@chromium.org chrome-secure-web-and-net@chromium.org Records the result of CheckCTRequirements when evaluating whether a SPDY or QUIC connection can pool with an existing session. mattm@chromium.org davidben@chromium.org Recorded each time a platform certificate change notification occurs. (Note that on Linux and ChromeOS these events are generated by the Chrome Certificate Manager but there are no notifications for changes that occurred outside of Chrome, and on Windows there are no notifications at all.) This does not count internal changes like Chrome Root Store and CRLSet updates. mattm@chromium.org chrome-secure-web-and-net@chromium.org The SHA-256 hash of the subjectPublicKeyInfo of the most-specific trust anchor encountered in a successfully-verified certificate chain, during an HTTP URL request. The associated root certificate can be looked up using https://crt.sh/?spkisha256= . This histogram is being replaced by Net.Certificate.TrustAnchor2.Request and will be removed once the new histogram has built up enough history. mattm@chromium.org chrome-secure-web-and-net@chromium.org The SHA-256 hash of the subjectPublicKeyInfo of the most-specific trust anchor encountered in a successfully-verified certificate chain, during verification time. The associated root certificate can be looked up using https://crt.sh/?spkisha256= . This histogram is being replaced by Net.Certificate.TrustAnchor2.Verify and will be removed once the new histogram has built up enough history. mattm@chromium.org chrome-secure-web-and-net@chromium.org This histogram is recorded every time Chrome successfully validates a certificate, but cannot determine a public trust anchor from the verified chain. A value of true indicates that the trust anchor list is out of date, as the OS believes a public trust anchor is being used, but the built-in list does not know about this trust anchor, while a value of false indicates that a public trust anchor was not used, thus there is no evidence that the list is out of date. This histogram is temporary until the transition to a unified root store (in source), at which point, the OS store will no longer be consulted for known roots. mattm@chromium.org chrome-secure-web-and-net@chromium.org For each HTTP URL request satisfied by an active connection (non-cached requests), where the server certificate verified successfully and the Chrome Root Store is being used, records the Chrome Root Store stable ID of the root of trust of the successfully verified certificate chain. In the case where the Chrome Root Store includes trusted intermediate CA certificates and the verified path terminated at a trusted intermediate, the ID will be that of the root certificate, as if the path had been built all the way to the root. mattm@chromium.org chrome-secure-web-and-net@chromium.org For each certificate verification, where the certificate verified successfully and the Chrome Root Store is being used, records the Chrome Root Store stable ID of the root of trust of the successfully verified certificate chain. Does not include cached verifications, and is only recorded once when multiple identical verification inputs are coalesced into a single verification attempt. In the case where the Chrome Root Store includes trusted intermediate CA certificates and the verified path terminated at a trusted intermediate, the ID will be that of the root certificate, as if the path had been built all the way to the root. estark@chromium.org trusty-transport@chromium.org The origin breakdown of Signed Certificate Timestamps (SCTs). Emitted once for every SCT when first validated, which means 0 or more times during every SSL connection establishment. hchao@chromium.org mattm@chromium.org Time needed to initialize the Android platform Trust store when using the builtin certificate verifier on Android. mattm@chromium.org chrome-secure-web-and-net@chromium.org Records the result of CheckCTRequirements after a certificate verification. nidhijaju@chromium.org chrome-network-stack@google.com Records the time taken to verify the certificate on a worker thread for connections to {HostType}. This is recorded after a certificate is verified in the MultithreadedCertVerifier, and does not include the thread hop from the network service. mattm@chromium.org hchao@chromium.org When initializing the builtin certificate verifier on Mac using the DomainCacheFullCerts implementation of the system trust store integration, this records how long it took to query the Mac keychain and initialize the cache of intermediate certificates. Warning: this histogram was expired from 2023-04-09 to 2023-06-07(M116); data may be missing. mattm@chromium.org hchao@chromium.org When initializing the builtin certificate verifier on Mac using the DomainCacheFullCerts or KeychainCacheFullCerts implementation of the system trust store integration, this records the number of certificates that were included in the cache of intermediate certificates. mattm@chromium.org hchao@chromium.org When initializing the builtin certificate verifier on Mac using the DomainCacheFullCerts or KeychainCacheFullCerts implementation of the system trust store integration, this records the number of certificates found in the Mac keychain. mattm@chromium.org hchao@chromium.org When initializing the builtin certificate verifier on Mac using the KeychainCacheFullCerts implementation of the system trust store integration, this records the number of certificates found in the Mac keychain which had trust records marking them as either trusted or distrusted. mattm@chromium.org hchao@chromium.org When the builtin certificate verifier is used on Mac and the platform trusted certificates cache is initialized (or re-initialized if keychain trust settings changed), records how long it took to initialize all the trust domain caches. Only recorded by the DomainCacheFullCerts implementation, which may or may not be used depending on Finch trial parameters. mattm@chromium.org hchao@chromium.org When the builtin certificate verifier is used on Mac and the platform trusted certificates cache is initialized (or re-initialized if keychain trust settings changed), records the number of certificates read from the {Domain} trust domain. Only counts certificates that were successfully parsed. Only recorded by the DomainCacheFullCerts implementation, which may or may not be used depending on Finch trial parameters. Warning: this histogram was expired from 2023-04-09 to 2023-06-07(M116); data may be missing. mattm@chromium.org hchao@chromium.org When the builtin certificate verifier is used on Mac and the platform trusted certificates cache is initialized (or re-initialized if keychain trust settings changed), records how long it took to initialize both the trust cache and the intermediates cache. Only recorded by the DomainCacheFullCerts and KeychainCacheFullCerts implementations, which may or may not be used depending on Finch trial parameters. mattm@chromium.org rsleevi@chromium.org Whether Name normalization was used in successfully validated certificate chains, only recorded for chains ending in privately trusted roots. Suffixed by the CertVerifyProc implementation. {CertVerifyProcImpl} mattm@chromium.org hchao@chromium.org Records the result of migrating certificates from the user's NSS database into the ServerCertificateDatabase. This only applies on ChromeOS when then kEnableCertManagementUIV2Write flag is enabled, and is only recorded when migration is required (once per profile, on the first time the profile is loaded after the feature flag is flipped to true). mattm@chromium.org src/net/cert/OWNERS Result of attempting to verify if a certificate is a valid 1-QWAC (EU qualified website authentication certificate). This will be emitted only when the VerifyQWACs feature flag is enabled, Chrome Root Store is being used and CertVerifyProcBuiltin::VerifyInternal returns OK. mattm@chromium.org src/net/cert/OWNERS Result of attempting to verify if a 2-QWAC (EU qualified website authentication certificate) binding is valid and binds the specified TLS certificate. This will be emitted only when the VerifyQWACs feature flag is enabled, Chrome Root Store is being used, the site's TLS certificate validates successfully, the request includes a 2-QWAC link header, and fetching and processing the 2-QWAC binding is successful. mattm@chromium.org src/net/cert/OWNERS Result of processing 2-QWAC link headers. This will be emitted only for committed TLS main frame navigations when the VerifyQWACs feature flag is enabled and Chrome Root Store is being used. (Notably, it is not recorded for same-page navigations, nor are Link headers on main-frame redirects processed or histogrammed.) mattm@chromium.org rsleevi@chromium.org The actual amount of time spent verifying a certificate using the underlying cryptographic APIs. Because parallel verifications for the same certificate may be coalesced, histograms such as Net.SSLCertVerificationTime may be skewed, due to later verifications taking less overall time. This records the overall time spent verifying the first job to capture initialization costs. rsleevi@chromium.org mattm@chromium.org The actual amount of time spent verifying a certificate using the underlying cryptographic APIs. Because parallel verifications for the same certificate may be coalesced, histograms such as Net.SSLCertVerificationTime may be skewed, due to later verifications taking less overall time. This records the overall time spent verifying a single request, regardless of how many parallel requests are being served by the verification. dschinazi@chromium.org src/net/OWNERS Application protocol used for main frame resources. Logged by the renderer only for request that accessed the network and have received response headers. dschinazi@chromium.org src/net/OWNERS Application protocol used for subresources (resources other than main frame). Logged by the renderer only for request that accessed the network and have received response headers. nidhijaju@chromium.org blink-network-stack@google.com For each response stream, record the Content-Encoding type. If multiple encoding types are specified, this histogram only records the last one. Note: Compression Dictionary Transport headers (i.e. "sbr" for Shared Brotli) will be recorded as "Unknown" in this histogram. bashi@chromium.org src/net/OWNERS The number of broken alternative services loaded from the prefs file on startup. Alternative services allow an origin server to specify additional means of interacting with it on the network. A broken alternative service is one whose most recent connection attempt failed and whose retry timeout period has not yet expired. Recorded when broken alternative services are loaded from the prefs file on startup. dschinazi@chromium.org src/net/quic/OWNERS This counts the number of server configs persisted in prefs file. dschinazi@chromium.org src/net/OWNERS The number of recently broken alternative services loaded from the prefs file on startup. Alternative services allow an origin server to specify additional means of interacting with it on the network. A recently broken alternative service is one whose most recent connection attempt failed. Recorded when broken alternative services are loaded from the prefs file on startup. jdoerrie@chromium.org chrome-counter-abuse-alerts@google.com This records the outcome of attempting to save the hardware attestation key during a Device Bound Session store operation (saving a new session or updating an existing one). drubery@chromium.org chrome-counter-abuse-alerts@google.com This records the outcome of deferring a request due to a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md). This is logged every time a {Outcome} request is unblocked after being deferred, which is usually due to a refresh completing. drubery@chromium.org chrome-counter-abuse-alerts@google.com This records the outcome of deferring a request due to a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md). This is logged every time a request is unblocked after being deferred {Duration}, which is usually due to a refresh completing. drubery@chromium.org chrome-counter-abuse-alerts@google.com This records the reason a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md) was deleted. This is logged every time we terminate a session. thefrog@chromium.org chrome-counter-abuse-alerts@google.com This records whether a Google Device Bound Session registration was performed by the Device Bound Session Standard in Net code. The alternative is that it was done by BoundSessionCredentials SignIn code. Standard Device Bound Session registration is described in (https://github.com/w3c/webappsec-dbsc/blob/main/README.md). This histogram is logged every time the Standard or the BoundSessionCredentials SignIn code has just completed trying to register a new Google session, whether or not that registration succeeded. drubery@chromium.org chrome-counter-abuse-alerts@google.com On each refresh request due to a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md), this histogram logs the minimum bound cookie lifetime in the refresh response. drubery@chromium.org chrome-counter-abuse-alerts@google.com On each refresh request due to a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md), this histogram logs the minimum threshold to make the previous request trigger a proactive refresh. This is logged {Outcome}. santosls@google.com chrome-counter-abuse-alerts@google.com Records the number of consecutive failures when the target_url provided to DeviceBoundSessionPrewarmer is invalid, empty, or not HTTPS, forcing it to fall back to a long prewarm interval. Recorded when DoPrewarm runs. santosls@google.com chrome-counter-abuse-alerts@google.com Records the outcome of a prewarm operation per DBSC session. This is logged to {PrewarmType} track. Startup represents the first prewarm when started, and Scheduled represents the prewarms that are scheduled based on the Mojo response. drubery@chromium.org chrome-counter-abuse-alerts@google.com This records the outcome of attempting to proactively refresh a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md#refresh-procedure). It records the reasons we may not have attempted to refresh despite a bound cookie expiring soon. This is logged every time a request is made with a bound cookie expiring soon. For the final result of proactive refresh, see Net.DeviceBoundSessions.RefreshResult.Proactive. drubery@chromium.org chrome-counter-abuse-alerts@google.com On each refresh request due to a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md) that Chrome believes may have been preventable with proactive refresh, this histogram logs the number of deferrals that proactive refresh would have prevented. This is logged when the minimum threshold is {Threshold}. drubery@chromium.org chrome-counter-abuse-alerts@google.com On each refresh request due to a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md) that Chrome believes may have been preventable with proactive refresh, this histogram logs the duration of each deferral that proactive refresh would have prevented. This is logged when the minimum threshold is {Threshold}. drubery@chromium.org chrome-counter-abuse-alerts@google.com This records the duration of a proactive refresh of a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md#refresh-procedure). This is logged every time a proactive refresh is done, which is triggered by a request with a bound cookie expiring soon. alexilin@chromium.org ahijazi@chromium.org chrome-counter-abuse-alerts@google.com This records the network result of a refresh request for a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md). This is logged when refresh network response processing finishes for {Attempt}. If a network error occurs at any point while reading the response body, that is logged instead of the HTTP status code. drubery@chromium.org chrome-counter-abuse-alerts@google.com This records the number of challenge signings done during one Device Bound Session refresh (https://github.com/w3c/webappsec-dbsc/blob/main/README.md). This is logged at the end of each session refresh. alexilin@chromium.org chrome-counter-abuse-alerts@google.com This records the last outcome of refreshing a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md#refresh-procedure) when we reject a refresh due to exceeding the refresh quota. drubery@chromium.org chrome-counter-abuse-alerts@google.com This records the outcome of refreshing a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md#refresh-procedure). This is logged every time we attempt to refresh the cookies bound by a session due to {RefreshType}. drubery@chromium.org chrome-counter-abuse-alerts@google.com This records the outcome of {RegistrationType} Device Bound Session registration (https://github.com/w3c/webappsec-dbsc/blob/main/README.md). This is logged every time we attempt to register a new session. drubery@chromium.org chrome-counter-abuse-alerts@google.com This records how many times a request was deferred due to a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md). This is logged once for each round trip of a network requests that can add a cookie header; requests that attempt to refresh but fail are omitted. thefrog@chromium.org chrome-counter-abuse-alerts@google.com This records whether a request was deferred due to a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md) and the reason for that decision. This is logged once for each round trip of most network requests that can add a cookie header; requests that attempt to refresh but fail are omitted. drubery@chromium.org chrome-counter-abuse-alerts@google.com This records how many requests were deferred by a single Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md) refresh. This is logged every time requests are unblocked, which is usually due to a refresh completing. drubery@chromium.org chrome-counter-abuse-alerts@google.com This records the duration of each time a request is deferred due to a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md). This is logged every time one round trip of a network request is deferred. Under normal operations, a small fraction of network requests are deferred at most once. See also Net.DeviceBoundSessions.TotalRequestDeferredDuration. drubery@chromium.org chrome-counter-abuse-alerts@google.com This records the number of network requests deferred while we waited to load Device Bound Sessions (https://github.com/w3c/webappsec-dbsc/blob/main/README.md) persisted to disk. It is logged once at startup, when sessions are loaded. drubery@chromium.org chrome-counter-abuse-alerts@google.com This records how long it took to load Device Bound Sessions (https://github.com/w3c/webappsec-dbsc/blob/main/README.md) persisted to disk. It is logged once at startup, when sessions are loaded. drubery@chromium.org chrome-counter-abuse-alerts@google.com This records whether loading existing Device Bound Sessions (https://github.com/w3c/webappsec-dbsc/blob/main/README.md) that were persisted to disk succeeded. It is logged once at startup, when sessions are loaded. thefrog@chromium.org chrome-counter-abuse-alerts@google.com This records the last outcome of refreshing a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md#refresh-procedure) when we reject a signing due to exceeding the signing quota. drubery@chromium.org chrome-counter-abuse-alerts@google.com This records the total amount of time a request spent deferred due to a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md). This is logged once for every network request that is deferred, totalling the delay due to all deferrals. Under normal operations, this is a small fraction of network requests. thefrog@chromium.org chrome-counter-abuse-alerts@google.com Logs whether there is already a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md) {RefreshType} refresh attempt in progress, so there is no need to trigger a new refresh for this request. This histogram is logged when the Standard is attempting to defer a request for a valid session's refresh. thefrog@chromium.org chrome-counter-abuse-alerts@google.com This records the network result of a {RequestType} request for a Device Bound Session (https://github.com/w3c/webappsec-dbsc/blob/main/README.md). This is logged when {RequestType} network response processing finishes. If a network error occurs at any point while reading the response body, that is logged instead of the HTTP status code. ricea@chromium.org src/net/OWNERS Records the number of socket connect calls made during a single call to net::AddressSorterPosix::Sort(). Recorded once for every call to Sort(). Only recorded on platforms that use the Posix implementation. ricea@chromium.org src/net/OWNERS Records the time it takes to execute a single call to net::AddressSorterPosix::Sort(). Measured for every call to Sort(). Only recorded on platforms that use the Posix implementation. The histogram is not reported on systems without high resolution clocks. horo@chromium.org src/net/OWNERS Whether or not `prefs::kAdditionalDnsQueryTypesEnabled` is enabled. Recorded at startup if disabled or whenever Chrome is believed to be enterprise-managed (per the same detection as the default disabling of DoH). Measures whether or not enterprise admins are setting the disable, potentially due to issues encoutered with HTTPS/INTEGRITY query experiments. horo@chromium.org src/net/OWNERS Whether DNS clients can and are allowed to perform secure and insecure DNS transactions. Logged when HostResolverManager receives a host resolution request using HostResolverSource::ANY. horo@chromium.org src/net/OWNERS Whether or not the DNS configuration read from a nsswitch.conf file is considered compatible with Chrome, that is that the configuration did not contain any options that Chrome is unable to support using the built-in resolver. Reported whenever a nsswitch.conf file is read by Chrome. horo@chromium.org src/net/OWNERS Enumeration of the nsswitch.conf service that caused a compatibility issue, leading to the nsswitch.conf DNS configuration being considered incompatible with Chrome. Only emitted when a specific token is present in the nsswitch.conf file, causing the incompatibility, e.g. because the service is always incompatible with Chrome or because the service is configured with incompatible actions. horo@chromium.org src/net/OWNERS Whether or not an attempt to read nsswitch.conf succeeded. May fail due to missing file or filesystem issues, but does not consider it a failure if the file is larger than the size limit (see Net.DNS.DnsConfig.Nsswitch.TooLarge for that case). Reported for all attempts to read nsswitch.conf. horo@chromium.org src/net/OWNERS Whether or not an attempt to read nsswitch.conf failed due to the file size being unreasonably large. Reported for all attempts to read nsswitch.conf. horo@chromium.org src/net/OWNERS Whether or not the DNS configuration read from a resolv.conf file is considered compatible with Chrome, that is that the configuration did not contain any options that Chrome is unable to support using the built-in resolver. Reported whenever a resolv.conf file is successfully read by Chrome. horo@chromium.org src/net/dns/OWNERS Records details of the secure DNS mode used to configure the host resolver at construction of the network service. Warning: this histogram was expired from M140 to M142; data may be missing. horo@chromium.org src/net/OWNERS Enumeration of whether or not the DNS configuration setting of Windows is considered compatible with Chrome and the reason why it is considered incompatible. Reported whenever the DNS configuration setting of Windows is read by Chrome. horo@chromium.org src/net/OWNERS The result of reading system DNS settings. Reported whenever Chrome tries to read the DNS configuration setting of Windows. horo@chromium.org net-dev@chromium.org The number of hosts parsed from a Hosts file. Reported whenever a Hosts file is read and parsed. horo@chromium.org net-dev@chromium.org Estimated memory usage to keep the parsed result of a Hosts file. Reported whenever a Hosts file is read and parsed. horo@chromium.org net-dev@chromium.org The file size of a Hosts file. Reported whenever a Hosts file is read. stefanoduo@chromium.org src/net/dns/OWNERS Whether the file descriptor returned by android_res_nquery was already readable as soon as it was returned. Reported when starting a platform DNS attempt on Android. stefanoduo@chromium.org src/net/dns/OWNERS The size of the DNS response in bytes returned by android_res_nresult for {QueryType} resource records. Reported when successfully reading a platform DNS attempt response on Android. awillia@chromium.org src/net/dns/OWNERS Status of HTTPS DNS records received as additional records in a DnsTask querying for HTTPS. This would represent a recurssive resolver that is HTTPS-aware and performing followup queries on behalf of the client. Recorded as DNS responses are parsed. horo@chromium.org src/net/OWNERS Enumeration of whether or not an HTTPS (SVCB) record transaction succeeded and whether or not any failure is fatal to the DnsTask. Recorded for all completed HTTPS record transactions. bashi@chromium.org net-dev@chromium.org Duration of time taken by a {Secure}DnsTask (the grouping of all individual DNS queries for a single instance of resolving a name) in resolutions that failed. Recorded when a {Secure}DnsTask failed. bashi@chromium.org net-dev@chromium.org Duration of time taken by a {Secure}DnsTask (the grouping of all individual DNS queries for a single instance of resolving a name) in resolutions that succeeded. Recorded when a {Secure}DnsTask succeeded. horo@chromium.org src/net/OWNERS Enumeration of the logic path used (e.g. different protocols) for each started DnsAttempt (the representation of an individual network request to a DNS server). awillia@chromium.org src/net/dns/OWNERS Duration of time taken by an insecure DnsTransaction to fail. {DohProviderId} awillia@chromium.org src/net/dns/OWNERS Duration of time taken by an insecure DnsTransaction to succeed. {DohProviderId} awillia@chromium.org src/net/dns/OWNERS Counts of specific error codes returned by secure DnsTransactions to a DoH server that is unavailable or has unknown availability. These transactions may be probe queries or queries issued in the secure DNS mode. {DohProviderId} awillia@chromium.org src/net/dns/OWNERS Counts of specific error codes returned by secure DnsTransactions to an available DoH server. A DoH server is available if a probe has been successfully completed and fewer than a threshold amount of failures have occurred. {DohProviderId} awillia@chromium.org src/net/dns/OWNERS Duration of time taken by a secure DnsTransaction to an available DoH server to fail. A DoH server is available if a probe has been successfully completed and fewer than a threshold amount of failures have occurred. {DohProviderId} awillia@chromium.org src/net/dns/OWNERS Duration of time taken by a secure DnsTransaction to an available DoH server to succeed. A DoH server is available if a probe has been successfully completed and fewer than a threshold amount of failures have occurred. {DohProviderId} horo@chromium.org src/net/OWNERS Enumerated status of the reason why the built-in DNS resolver has detected a "low entropy" situation for DNS over UDP and has switched to DNS over TCP. bashi@chromium.org blink-network-stack@google.com Measures whether a DoH request used a new or an existing session. Recorded when a non-probe, auto-upgrade mode DoH transaction completes. bashi@chromium.org blink-network-stack@google.com Measures whether early data was accepted for a DoH request. Recorded when a non-probe, auto-upgrade mode DoH transaction succeeds. bashi@chromium.org blink-network-stack@google.com The time taken for a SPDY/QUIC session to create an active stream for a DoH request. Measures pending time due to max stream limits. Recorded when a non-probe, auto-upgrade mode DoH transaction succeeds. bashi@chromium.org blink-network-stack@google.com Duration of time taken by a DoH transaction which {EarlyDataStatus}. Recorded when a non-probe, auto-upgrade mode DoH transaction succeeds. Valid only for requests that access the network. bashi@chromium.org blink-network-stack@google.com Duration of time taken by a DoH transaction to {Outcome}. Recorded when a non-probe, auto-upgrade mode DoH transaction completes. stefanoduo@google.com src/net/dns/OWNERS Net error codes resulting from completed or aborted DnsTransactions querying for {QueryType} resource records. Recorded across all resolution modes when a transaction completes or is aborted before completion. stefanoduo@google.com src/net/dns/OWNERS Net error codes resulting from completed or aborted classic (Do53) DnsTransactions querying for {QueryType} resource records that experienced UDP truncation (TC bit set) and retried over TCP. Recorded when a classic DNS transaction completes or is aborted after retrying over TCP. stefanoduo@google.com src/net/dns/OWNERS Net error codes resulting from completed or aborted DnsTransactions querying for {QueryType} resource records using {AttemptMode}. Recorded when a transaction completes or is aborted before completion. bcb@chromium.org src/net/dns/OWNERS Records the reason for falling back from a secure DNS transaction to an insecure one. This is recorded every time HostResolverManager::PushDnsTasks is called while in AUTOMATIC secure DNS mode. A preferred reason indicates that the SECURE_DNS task was prevented from being included to resolve a host request. suzukikeita@chromium.org blink-network-stack@google.com When erasing a stale (expired or old-network) DNS entry for {HostType} from the host cache, how long past the expiration time it is. This is recorded each time an erase happens, when clearing the host cache or evicting the entry. suzukikeita@chromium.org blink-network-stack@google.com When erasing a stale (expired or old-network) DNS entry for {HostType} from the host cache how many network changes happened between setting and erasing it. This is recorded each time an erase happens, when clearing the host cache or evicting the entry. WARNING: This metric was expired on 2025-10-26, but was re-enabled on 2026-01-20. Some data may be missing. suzukikeita@chromium.org blink-network-stack@google.com When erasing a stale (expired or old-network) DNS entry for {HostType} from the host cache how many hits it received while stale. This is recorded each time an erase happens, when clearing the host cache or evicting the entry. suzukikeita@chromium.org blink-network-stack@google.com When erasing a valid DNS entry for {HostType} from the host cache, for how much longer it would have remained valid. This is recorded each time an erase happens, when clearing the host cache or evicting the entry. WARNING: This metric was expired on 2025-10-26, but was re-enabled on 2026-01-20. Some data may be missing. suzukikeita@chromium.org blink-network-stack@google.com The reason for erasing a DNS entry for {HostType} from the host cache. This is recorded each time an erase happens, when clearing the host cache or evicting the entry. WARNING: This metric was expired on 2025-12-14, but was re-enabled on 2026-01-20. Some data may be missing. suzukikeita@chromium.org blink-network-stack@google.com When looking up a stale DNS entry for {HostType} in the host cache, how long past the expiration time it is. This is recorded at the end of each DNS entry lookup. WARNING: This metric was expired on 2025-06-10, but was re-enabled on 2026-01-20. Some data may be missing. This histogram has been expired between 2026-06-10 and 2026-07-01, thus does not contain data for this period. suzukikeita@chromium.org blink-network-stack@google.com When looking up a stale DNS entry for {HostType} in the host cache, how many network changes happened between setting it and looking it up. This is recorded at the end of each DNS entry lookup. WARNING: This metric was expired on 2025-08-10, but was re-enabled on 2026-01-20. Some data may be missing. suzukikeita@chromium.org blink-network-stack@google.com The outcome of looking up a DNS entry for {HostType} in the host cache. This is recorded at the end of each DNS entry lookup. bashi@chromium.org davidben@chromium.org src/net/OWNERS Records which RCODE the DNS server specified when it responded to our HTTPS query. If the query timed out before getting a response, "TimedOut" is recorded instead of an RCODE value. Histograms are emitted at the completion of a DnsTask's final transaction. bashi@chromium.org davidben@chromium.org src/net/OWNERS Indicates whether the result of an HTTPS query was validly parsable or mangled. Only recorded when DnsRcode indicates NoError. Histograms are emitted at the completion of a DnsTask's final transaction. horo@chromium.org davidben@chromium.org src/net/OWNERS Records true when we receive an HTTPS record and an error RCODE at the same time. Only recorded for true. Histograms are emitted at the completion of a DnsTask's final transaction. horo@chromium.org davidben@chromium.org src/net/OWNERS Records how many ms the DNS server took to respond to address queries. Histograms are emitted at the completion of a DnsTask's final transaction. horo@chromium.org davidben@chromium.org src/net/OWNERS Records how many ms the DNS server took to respond to an HTTPS query. Histograms are emitted at the completion of a DnsTask's final transaction. horo@chromium.org davidben@chromium.org src/net/OWNERS Records a ratio indicating the relative HTTPS query resolve time compared to the slower A or AAAA resolve time, where A/AAAA/HTTPS belong to the same DnsTask. Multiply this value by 10 to obtain a percentage P. The HTTPS query completed in roughly P% of the slower A/AAAA resolve time. Values smaller than 10 indicate that HTTPS query was faster than A/AAAA. Values larger than 10 indicate that the HTTPS query was slower than A/AAAA. Histograms are emitted at the completion of a DnsTask's final transaction. bashi@chromium.org net-dev@chromium.org Record whether a HostResolverManager::Job attempted HTTPS query. Recorded when the Job succeeded, had non speculative requests, and source type is ANY (which means that the Job could have queried more than one query). bashi@chromium.org net-dev@chromium.org Record why a HostResolverManager::Job didn't attempt HTTPS query. Recorded when the Job succeeded, had non speculative requests, source type is ANY (which means that the Job could have queried more than one query), and didn't attempt HTTPS query. bashi@chromium.org net-dev@chromium.org The duration of time DnsTransactions spend waiting to be started by a DnsTask, aggregated by HostResolverManager::Job. This metric is only recorded when the Job failed. bashi@chromium.org net-dev@chromium.org The duration of time a DnsTransaction spends waiting to be started by a DnsTask. The dominant factor is likely the time spent waiting for other DnsTransactions to complete, i.e. waiting for DnsTask::StartNextTransaction to be invoked by the dispatcher. If there were no limit on the number of concurrent DNS queries, we would expect this metric to be near zero. bashi@chromium.org net-dev@chromium.org The duration of time DnsTransactions spend waiting to be started by a DnsTask, aggregated by HostResolverManager::Job. This metric is only recorded when the Job succeeded. awillia@chromium.org src/net/dns/OWNERS Duration of time taken by a probe attempt to fail. The time is measured since the probe sequence was initiated due to a DNS config change. awillia@chromium.org src/net/dns/OWNERS Duration of time taken by a probe attempt to succeed. The time is measured since the probe sequence was initiated due to a DNS config change. awillia@chromium.org src/net/dns/OWNERS Duration of time taken by a probe attempt to fail. The time is measured since the probe sequence was initiated due to a network change. awillia@chromium.org src/net/dns/OWNERS Duration of time taken by a probe attempt to succeed. The time is measured since the probe sequence was initiated due to a network change. horo@chromium.org src/net/OWNERS Duration of HostResolver requests from Start() until completion. Includes requests resolved synchronously using fast local sources, e.g. from cache, recorded as 0. Excludes speculative requests. horo@chromium.org src/net/OWNERS Duration of HostResolver requests from Start() until completion. Excludes speculative requests and requests resolved using fast local sources, e.g. from cache. horo@chromium.org src/net/OWNERS Whether a HostResolverManager::Job (the basic internal unit of work for non-local host resolutions, potentially merging redundant requests from Chrome) succeeded or failed, and whether it was speculative. awillia@chromium.org src/net/OWNERS The state of the Secure DNS (DoH) auto-upgrade functionality. This histogram is only emitted by clients with secure DNS set to automatic mode, and the emission occurs one minute after the insecure DNS to secure DNS auto-upgrade network probing begins. There will be one histogram emission per DoH provider in the system DNS configuration per NetworkContext that allows using secure DNS queries, and re-emissions can occur each time the DNS configuration changes. {DohProviderId} horo@chromium.org src/net/OWNERS For DNS resolutions that failed after less than 10 ms, which are probably local failures, what error code the jobs failed with. horo@chromium.org src/net/OWNERS For DNS resolutions that failed after at least 10 ms, which are probably remote failures or connectivity problems, what error code the jobs failed with. horo@chromium.org src/net/OWNERS Duration of time taken by HostResolverManager::Job (the basic internal unit of work for non-local host resolutions, potentially merging redundant requests from Chrome) in resolutions that failed. This is the time to resolve a hostname from start to finish. Excludes jobs where all requests are speculative. bashi@chromium.org net-dev@chromium.org The complete fallback path taken to achieve a successful host resolution. Emitted when a non-speculative HostResolverManager::Job successfully resolves a host. horo@chromium.org src/net/OWNERS Duration of time taken by HostResolverManager::Job (the basic internal unit of work for non-local host resolutions, potentially merging redundant requests from Chrome) in resolutions that succeeded. This is the time to resolve a hostname from start to finish. Excludes jobs where all requests are speculative. bashi@chromium.org chrome-network-stack@google.com Time difference between when DNS records were resolved. Recorded when {Variant}. horo@chromium.org src/net/OWNERS Duration of time taken by HostResolverSystemTask (the grouping of all individual DNS queries for a single instance of resolving a name using the system resolver) in resolutions that failed. horo@chromium.org src/net/OWNERS Duration of time taken by HostResolverSystemTask (the grouping of all individual DNS queries for a single instance of resolving a name using the system resolver) in resolutions that succeeded. awillia@chromium.org src/net/dns/OWNERS Records the outcome of a DoH probe attempt on the first valid entry in the custom text field. awillia@chromium.org src/net/dns/OWNERS Records the outcome of a formatting validation attempt on the custom text field. awillia@chromium.org src/net/dns/OWNERS True if upgrade to DoH from a DoT hostname was attempted and succeeded. False if it was attempted and failed. awillia@chromium.org src/net/dns/OWNERS True if there was at least one public nameserver during an attempt to upgrade to DoH from insecure DNS. False if there were no public nameservers during an attempt to upgrade to DoH from insecure DNS. Warning: this histogram was expired from 2023-04-23 to 2023-06-12; data may be missing. awillia@chromium.org src/net/dns/OWNERS True if upgrade to DoH was not attempted because the DoH config was already specified. False if upgrade to DoH was not attempted for some other reason. awillia@chromium.org src/net/dns/OWNERS True if upgrade to DoH was not attempted because of unhandled options in the system config. False if upgrade to DoH was not attempted for some other reason. Emitted each time the DnsConfig is updated. awillia@chromium.org src/net/dns/OWNERS True if upgrade to DoH from insecure DNS was attempted and succeeded. False if it was attempted and failed. kaklilu@chromium.org awillia@chromium.org src/net/dns/OWNERS Recorded if upgrading to DoH from insecure DNS was attempted, failed, then succeeded using the fallback DoH server config. This is only emitted when Secure DNS is used in automatic-with-doh-fallback mode. We also only try using the fallback if and the system DNS provider was not in Chrome's supported DoH provider list. kaklilu@chromium.org awillia@chromium.org src/net/dns/OWNERS Breaks down what kinds of local nameservers (if any) are on the DNS config when Chrome has tried and failed to upgrade it to DoH from insecure DNS. mmenke@chromium.org The time measured before starting DNS lookup until after the connection is complete. ricea@chromium.org src/net/network_error_logging/OWNERS The outcome of a Domain Reliability beacon, logged upon destruction of the beacon. arichiv@chromium.org awillia@chromium.org katabolism-finch@google.com Logged on connection completion when EphemeralPortRandomizer was used to pick the port. Records the associated net error (or OK). MacOS only. arichiv@chromium.org awillia@chromium.org katabolism-finch@google.com Logged as true when EphemeralPortRandomizer::PickPort fails to find a port to allocate for a pending TCP connection. Never logged as false. MacOS only. mmenke@chromium.org Positive net error codes that requests for pages end with, including net::OK and net::ERR_ABORTED. This only counts loads in "main frames" for https://www.google.com (it does not for example count the error codes for subresoures on a page). It excludes main frame navigations that are converted to downloads (unless they fail / are aborted before it's known that they would be a download), and requests handled externally (like mailto links). giovax@chromium.org pwa-commercial@google.com src/chrome/browser/web_applications/isolated_web_apps/OWNERS Positive net error codes that HTTP requests using the isolated-app: URL scheme end with, including net::OK and net::ERR_ABORTED. mmenke@chromium.org Positive net error codes that requests for pages end with, including net::OK and net::ERR_ABORTED. This only counts loads in "main frames" (it does not for example count the error codes for subresoures on a page). It excludes main frame navigations that are converted to downloads (unless they fail / are aborted before it's known that they would be a download), and requests handled externally (like mailto links). mmenke@chromium.org Net error codes that requests for "subresources" end with, including net::OK and net::ERR_ABORTED. edwardjung@chromium.org src/net/OWNERS Counts of various events that can occur on the network error page. Note that this histogram enum buckets are not mutually exclusive and so a single error page load might report samples into multiple ones. Each bucket has descriptions that further explain the context of their samples. finnur@chromium.org alexbn@chromium.org Counts the number of times web apps receive `ERR_INTERNET_DISCONNECTED` and the alternative error page (rather than the dino) is shown. Reported when the error page html is being prepared. See `NetErrorHelper::GenerateLocalizedErrorPage` (crbug.com/1280518). mpdenton@chromium.org src/net/OWNERS Per-authentication-scheme counts of HTTP authentication attempts and rejections. mpdenton@chromium.org src/net/OWNERS Per-authentication-scheme counts of HTTP authentication targets, such as secure servers or proxies. Emitted at the same time as Net.HttpAuthCount, on every HTTP authentication attempt or rejection. deepakr@chromium.org net-dev@chromium.org The number of active logical HTTP cache invalidation filters present in memory at the moment a new invalidation filter is added. Recorded in HttpCache::AddInvalidationFilter upon each filter addition. deepakr@chromium.org net-dev@chromium.org The context under which a logical HTTP cache invalidation filter is cleared (i.e., whether it was cleared during the same session it was added, or if it was successfully recovered from disk after a crash/restart and then subsequently cleared). Recorded when a filter is removed. deepakr@chromium.org net-dev@chromium.org Recorded as true when adding a new logical HTTP cache invalidation filter causes the oldest active filter to be evicted due to reaching the maximum filter capacity cap (configured via FeatureParam kLogicalClearHttpCacheMaxFilters, defaulting to 500). Recorded in HttpCache::AddInvalidationFilter. deepakr@chromium.org net-dev@chromium.org The number of logical HTTP cache invalidation filters successfully loaded from persistent disk storage at startup. Recorded once at NetworkContext startup when initializing the LogicalInvalidationStore. deepakr@chromium.org net-dev@chromium.org Records the result of loading the disk persistence file for logical HTTP cache invalidation filters (e.g. success, file not found, or corrupt). Recorded once at NetworkContext startup when initializing the LogicalInvalidationStore. deepakr@chromium.org net-dev@chromium.org The lookup match outcome when querying active HTTP cache invalidation filters (e.g. filter matched resulting in a cache miss, or filter did not match). Recorded on the hot path in HttpCache::IsInvalidated only when active invalidation filters are present in memory. deepakr@chromium.org net-dev@chromium.org The time taken to asynchronously load and deserialize logical HTTP cache invalidation filters from persistent disk storage via background thread. Recorded once at NetworkContext startup when initializing the LogicalInvalidationStore. deepakr@chromium.org net-dev@chromium.org The time taken to serialize and atomically write the active logical HTTP cache invalidation filters to persistent disk storage. Recorded every time LogicalInvalidationStore saves filters to disk. deepakr@chromium.org net-dev@chromium.org Recorded as true when a cache entry is matched and invalidated by a logical filter that was recovered from persistent storage after a crash or restart. This tracks how often recovered logical filters actively protect against serving stale data before a physical purge occurs. ashleynewson@chromium.org src/android_webview/OWNERS Total bytes received over the network for an HttpJob request. This is measured when the HttpJob is completed. Bytes after switching protocols (such as WebSockets) are not counted. There is no previous version of this metric. It is numbered to be in sync with similar metrics. Unlike PrefilterBytesRead, this metric accounts for all HTTP stream bytes (not just the content body), counts bytes for aborted jobs, and network bytes in 304 Not Modified exchanges (rather than cached content bodies). ashleynewson@chromium.org src/android_webview/OWNERS Total bytes sent over the network for an HttpJob request. This is measured when the HttpJob is completed. Bytes after switching protocols (such as WebSockets) are not counted. There is no previous version of this metric. It is numbered to be in sync with similar metrics. fayang@chromium.org src/net/quic/OWNERS Wait time of main job while there is an available SPDY session but no QUIC session. fayang@chromium.org src/net/quic/OWNERS Wait time of main job while there is no available SPDY session and no QUIC session. jkarlin@chromium.org chrome-ads-histograms@google.com Total prefilter (e.g., before decompression) bytes read for an HttpJob request that is run on behalf of a request tagged as advertising related. jkarlin@chromium.org shivanisha@chromium.org src/net/OWNERS Total prefilter (e.g., before decompression) bytes read for an HttpJob request. dschinazi@chromium.org rch@chromium.org src/net/OWNERS Scheme of the proxy server used by an HTTP job that failed, or SCHEME_INVALID if no proxy was used. A value is emitted for each failed HTTP stream. Note: Although SCHEME_DIRECT is deprecated, value 1 corresponds to a direct proxy. Warning: this histogram was expired from 2022-11-13 to 2023-12-01; data may be missing. dschinazi@chromium.org rch@chromium.org src/net/OWNERS Scheme of the proxy server used by an HTTP job that succeeded, or SCHEME_INVALID if no proxy was used. A value is emitted for each successful HTTP stream. Note: Although SCHEME_DIRECT is deprecated, value 1 corresponds to a direct proxy. Warning: this histogram was expired from 2023-03-19 to 2023-12-01; data may be missing. bashi@chromium.org trusty-transport@chromium.org Time it takes to complete an HttpJob to a Google host, from starting the transaction until we are done reading, if the response was received over TLS 1.3 and TCP. This is recorded to measure the impact of 0-RTT. mmenke@chromium.org Time it takes to complete an HttpJob, from starting the transaction until we are done reading, for jobs served from the cache. mmenke@chromium.org Time it takes to complete an HttpJob, from starting the transaction until the job is killed. Note that we didn't detect the end of the data for this job. mmenke@chromium.org Time it takes to complete an HttpJob, from starting the transaction until we are done reading, for jobs not served from the cache. {TotalTimeToHttpsGoogle} mmenke@chromium.org Time it takes to complete an HttpJob, from starting the transaction until we are done reading, for jobs when we read until no more data is available. {Priority} mmenke@chromium.org Time it takes to complete an HttpJob, from starting the transaction until we are done reading. {TotalTimeToHttpsGoogle} jbroman@chromium.org liviutinta@chromium.org Whether any unrecognized keys were seen at the root dictionary of a No-Vary-Search response header. Recorded each time one is parsed as a dictionary, whether or not it affects caching in any way. dschinazi@chromium.org src/net/OWNERS Time taken to establish the connection to an HTTP proxy. This is only measured for connections to the first proxy in a proxy chain. In the Error and TimedOut variations, HTTP/1 and HTTP/2 cannot be distinguished, so these are reported as HTTP/1. This histogram expired 2024-09-17 and was re-established in M131 with the addition of the HTTP version and proxy scheme. ricea@chromium.org net-dev@chromium.org Records macOS error codes returned by the macOS System Proxy Resolver when proxy resolution fails. This histogram is only recorded when a non-zero OS error code is returned, indicating a macOS API failure during proxy resolution. ricea@chromium.org net-dev@chromium.org Records the status of macOS system proxy resolution requests. This histogram is recorded for every proxy resolution request made through the macOS system proxy resolver, indicating whether the resolution succeeded or failed. Aborted requests (e.g., during browser shutdown) are not recorded. ricea@chromium.org src/services/proxy_resolver_win/OWNERS Logs the status of WinHTTP session initialization for proxy resolution on Windows. This entry is recorded when the WinHTTP session is initialized, which involves opening the session, configuring timeouts, and registering the status callback. ricea@chromium.org net-dev@chromium.org Records Windows error codes returned by the Windows System Proxy Resolver when proxy resolution fails. This histogram is only recorded when a non-zero Windows error code is returned, indicating a Windows API failure during proxy resolution. One of the error codes from https://learn.microsoft.com/en-us/windows/win32/com/com-error-codes-4. cthomp@chromium.org etienneb@chromium.org joenotcharles@google.com Recorded for all top-frame HTTP/HTTPS/WS/WSS requests made by UrlRequestHttpJob. Records whether the request was upgraded from HTTP (or WS) to HTTPS (or WSS), and whether the decision to upgrade required dynamic STS data. mmenke@chromium.org The count of HTTP Response codes encountered. nidhijaju@chromium.org ricea@chromium.org The count of HTTP Response codes encountered, for codes registered in the IANA HTTP Status Code Registry. Ignores unknown response codes. Recorded when HttpResponseHeaders are constructed, after parsing. bashi@chromium.org blink-network-stack@google.com Record a delay between an HttpStreamFactory::Job gets a successful stream and the job calls the delegate's OnStreamReady(). Recorded when the job gets a successful stream. bashi@chromium.org blink-network-stack@google.com Time from when an HttpStreamPool::AttemptManager is created to when the AttemptManager is destroyed. Recorded when an AttemptManager is destroyed. bashi@chromium.org blink-network-stack@google.com Time from when a service endpoint request is started to when an existing {Protocol} session is found by IP-based matching. Recorded when an existing {Protocol} session is found after the service endpoint request is updated or finished. bashi@chromium.org blink-network-stack@google.com The time taken from an HttpStreamPool::AttemptManager creation to the time the AttemptManager attempting the initial TCP based attempt. Recorded when an AttemptManager attempts TCP for the first time. bashi@chromium.org blink-network-stack@google.com Record the initial attempt state (whether known to support QUIC/SPDY, attempt QUIC first, etc). Recorded the first time a HttpStreamPool::AttemptManager attempts a TCP connection. bashi@chromium.org blink-network-stack@google.com Time from when an HttpStreamPool::Job is created to when the job is completed. Recorded for each job that has a result upon destruction of the job. bashi@chromium.org blink-network-stack@google.com Record a delay between an HttpStreamPool::JobController gets a successful stream and the job controller complets the corresponding request. Recorded when the job controller completes the corresponding request. bashi@chromium.org blink-network-stack@google.com Time from when an HttpStreamPool::JobController is created to when the job controller completes the corresponding request. Recorded when the job controller completes the corresponding request. bashi@chromium.org blink-network-stack@google.com The net error for an HttpStreamPool::Job. Recorded for each job that has a result upon destruction of the job. bashi@chromium.org blink-network-stack@google.com Time from when a stream attempt is created to when a new SPDY session is created from the attempt. Recorded when a new SPDY session is created as a result of a stream attempt. bashi@chromium.org blink-network-stack@google.com Time from when a QuicAttempt attempted a QUIC session to when the task is completed. Recorded for each QuicAttempt if the task actually attempted a session. bashi@chromium.org blink-network-stack@google.com Record the initial attempt state (whether known to support QUIC/SPDY, attempt QUIC first, etc) when a TcpBasedAttempt is canceled. Recorded for each stream attempt upon destruction of the canceled attempt. bashi@chromium.org blink-network-stack@google.com Time from when a stream attempt is created to when the attempt is canceled. Recorded for each stream attempt upon destruction of the canceled attempt. bashi@chromium.org blink-network-stack@google.com The reason why a stream attempt is canceled. Recorded when a stream attempt is destroyed after being canceled. bashi@chromium.org blink-network-stack@google.com Time to block TCP-based stream attempts. Recorded when an HttpStreamPool::AttemptManager is created. bashi@chromium.org blink-network-stack@google.com Time from when a TCP based attempt started waiting for service endpoint resolution to when the attempt gets the resolution. Recorded for each TCP based attempt upon destruction of the attempt if the attempt completed successfully. bashi@chromium.org blink-network-stack@google.com The number of TcpBasedAttemptSlots when an HttpStreamPool::AttemptManager cancels them. Recorded when an HttpStreamPool::AttemptManager cancels TcpBasedAttemptSlots. bashi@chromium.org blink-network-stack@google.com Time from when an HttpStreamPool::AttemptManager is started to when a TCP based attempt is started, measuring the delay of the TCP based attempt. Recorded for each TCP based attempt upon destruction of the attempt if the attempt completed successfully. bashi@chromium.org blink-network-stack@google.com Time from when a {Protocol} attempt is created to when the attempt is completed successfully. Recorded for each TCP based attempt upon destruction of the attempt if the attempt completed successfully. bashi@chromium.org blink-network-stack@google.com Time from when a TCP based attempt is created to when the attempt is completed. Recorded for each TCP based attempt upon destruction of the attempt. bashi@chromium.org blink-network-stack@google.com The number of all TcpBasedAttempts when an HttpStreamPool::AttemptManager cancels them. Recorded when an HttpStreamPool::AttemptManager cancels all TcpBasedAttempts in TcpBasedAttemptSlots. bashi@chromium.org blink-network-stack@google.com Time from when a {Protocol} attempt starts a TCP handshake to when the TCP handshake completes. Recorded for each {Protocol} attempt upon destruction of the attempt if the attempt completed successfully. bashi@chromium.org blink-network-stack@google.com Time from when a TLS attempt starts a TLS handshake to when the TLS handshake completes. Recorded for each TLS attempt upon destruction of the attempt if the attempt completed successfully. bashi@chromium.org blink-network-stack@google.com The number of HttpStreamPool::Jobs when an HttpStreamPool::AttemptManager cancels them. Recorded when an HttpStreamPool::AttemptManager cancels HttpStreamPool::Jobs. davidben@chromium.org hchao@chromium.org chrome-secure-web-and-net@chromium.org Time from when an HTTP request is issued to when the first byte is processed, if the response was not cached and the connection used the Server Padding extension. davidben@chromium.org hchao@chromium.org chrome-secure-web-and-net@chromium.org Time from when an HTTP request is issued to when the first byte is processed, if the response was not cached, the connection used the Server Padding extension, and the request did not reuse a connection. Recorded on URLRequestHttpJob::OnStartCompleted. davidben@chromium.org trusty-transport@chromium.org Time from when an HTTP request to a Google host is issued to when the first byte is processed, if the response was not cached and received over TLS 1.3 and TCP. This is recorded to measure the impact of 0-RTT. mmenke@chromium.org Time from when an HTTP request is issued to when the first byte is processed. {Prefetch} nidhijaju@chromium.org chrome-network-stack@google.com Records the time taken to post a task to the worker thread pool, verify a certificate for connections to {HostType}, and post a task back to the browser main thread in the MultiThreadedCertVerifier. Recorded after verifying the certificate and returning back to the browser main thread in the MultiThreadedCertVerifier. davidben@chromium.org src/third_party/boringssl/OWNERS Records, on startup, for ARM devices running Android, whether the kernel is missing AT_HWCAP2 from the ELF auxv and /proc/cpuinfo reports support for ARMv8 features. Those features should also be reported in AT_HWCAP2. This is to measure the workaround in https://crbug.com/boringssl/46. Warning: this histogram was expired from 2024-06-30 to 2026-01-20; data may be missing. mmenke@chromium.org net-dev@chromium.org The number of times that network state is polled before invoking ConnectionTypeObserver::OnConnectionTypeChanged on Windows. Each IPAddressChange event results repeatedly polling the network, with polls 1 second apart, and a subset of those result in sending notifications, depending on network state. This is because the change OS change notification often happens before the modified network connections can actually be observed. This records how many times the network state is polled before a particular connection type change notification is sent. Note that the NetworkChangeNotifier may continue to poll after sending a connection type change notification. The logic to report changes is fairly complicated. See logging code to better understand the values. fbeaufort@chromium.org src/net/OWNERS The new network connection cost any time a change to the device's active default network is observed. dschinazi@chromium.org src/net/OWNERS The new network connection type any time a change to the device's active default network is observed. ksakamoto@chromium.org webpackage-dev@chromium.org When Network Error Logging observes a completed request of signed exchange that might generate a report, what happens to it. NEL observes all requests, both successful and unsuccessful. It sends reports for successful and unsuccessful requests at specified sampling rates so that error rates can be calculated. Insecure requests are discarded for having an insecure origin regardless of existence of a policy for the origin. hayato@chromium.org net-dev@chromium.org Records the size of the Network Error Logging (NEL) Service's task backlog queue. This is recorded when commands are issued to the NEL service while it is not yet initialized, causing the task to be queued in the backlog. This metric is sampled with a probability of 0.001 to reduce overhead. horo@chromium.org hayato@chromium.org src/net/OWNERS Recorded when the mitigation for priority starvation (switching to async retry after a threshold of connection errors) is triggered. {Occurrence} This is for {Protocol}. hayato@chromium.org blink-network-stack@google.com Measures the time taken to call http network transaction connected callback. Recorded every time the connected callback is called and completed. Emitted when the callback is completed. bashi@chromium.org blink-network-stack@google.com Deprecated. Replaced by Create{StreamType}Time4. bashi@chromium.org blink-network-stack@google.com The time taken for creating a(n) {StreamType}. Recorded in HttpNetworkTransaction::DoCreateStreamComplete when the stream creation completes successfully for the first time (i.e. not for retries). ricea@chromium.org src/net/OWNERS When an HTTP(S) request with a GET method is sent to the network within 10 seconds of another GET request with an identical URL, records the time interval between the first request and the second. Recorded once per duplicate request, but only when the "LogDuplicateRequests" feature is enabled. hayato@chromium.org src/net/OWNERS Recorded when the mitigation for priority starvation (switching to async retry after a threshold of connection errors) is triggered. Records the initial network error code that started the connection error retry loop. This is for {Protocol}. bashi@chromium.org src/net/OWNERS Records whether HTTP stream initialization is blocked or not. Recorded every time HttpStream::InitializeStream() is called. bashi@chromium.org src/net/OWNERS The time taken for HTTP stream initialization to finish if the initialization was blocked. Recorded in HttpNetworkTransaction::DoInitStreamComplete when the stream initialization is complete. bashi@chromium.org src/net/OWNERS Deprecated. Replaced by NegotiatedProtocol3. bashi@chromium.org src/net/OWNERS Negotiated protocol of an HttpStream. Recorded in HttpNetworkTransaction when an HttpStreamRequest succeeded for the first time (i.e. not for retries). bashi@chromium.org blink-network-stack@google.com Measures the CPU time taken by creating a HttpStreamRequest. Recoreded every time the HttpNetworkTransaction::DoCreateStream() is called for HTTP stream request. horo@chromium.org hayato@chromium.org src/net/OWNERS The number of retry attempts made for connection errors during the lifetime of a single HttpNetworkTransaction. Recorded when the transaction is destroyed, if there was at least one retry attempt. The maximum value is kMaxRetryAttemptsOnConnectionErrors (50). This is for {Protocol}. hayato@chromium.org src/net/OWNERS Recorded when a retry attempt is made for a connection error during an HTTP network transaction. Records the specific network error code that triggered the retry. This is for {Protocol}. bashi@chromium.org blink-network-stack@google.com Deprecated. Replaced by SessionSource4. bashi@chromium.org blink-network-stack@google.com Measures whether a request used a new or an existing {Protocol} session. Recorded in HttpNetworkTransaction::DoCreateStreamComplete when the stream creation completes successfully for the first time (i.e. not for retries). Note that this is only recorded for HTTP/2 and HTTP/3. There are similar histograms called PageLoad.Clients.GoogleSearch.SessionSource.{Protocol} that are recorded only for search navigation. These histograms are recorded per navigation, whereas this histograms measures per request. bashi@chromium.org src/net/OWNERS Deprecated. Replaced by StreamAddressFamily3. bashi@chromium.org src/net/OWNERS Address family of a remote endpoint of an HttpStream. Recorded in HttpNetworkTransaction when an HttpStreamRequest succeeded and getting the remote endpoint IP address succeeded for the first time (i.e. not for retries). bashi@chromium.org src/net/OWNERS Deprecated. Replaced by StreamRequestCompleteTime4. bashi@chromium.org src/net/OWNERS The time taken by an HttpStreamRequest to complete. Recorded only when an HttpStreamRequest succeeded without retries in HttpNetworkTransaction for the first time (i.e. not for retries). bashi@chromium.org src/net/OWNERS Deprecated. Replaced by StreamRequestErrorCode4. bashi@chromium.org src/net/OWNERS The net error for an HttpStreamRequest. Recorded in HttpNetworkTransaction when an HttpStreamRequest failed for the first time (i.e. not for retries). hayato@chromium.org src/net/OWNERS Recorded when an HTTP network transaction has reached the maximum number of retries for connection errors. Recorded every time when the retry reached kMaxRetryAttemptsOnConnectionErrors (50). This is for {Protocol}. bashi@chromium.org blink-network-stack@google.com Deprecated. Replaced by {Protocol}StreamCreationTime3. bashi@chromium.org blink-network-stack@google.com The time taken by an HttpNetworkTransaction to create a successful HttpStream, recorded separately by protocol and whether the stream is created on top of new or existing socket/session. Recorded in HttpNetworkTransaction::DoCreateStreamComplete when the stream creation completes successfully for the first time (i.e. not for retries). horo@chromium.org src/net/OWNERS Log the reason when HttpNetworkTransaction retries. dschinazi@chromium.org src/net/quic/OWNERS The number of QUIC sessions when the QuicSessionPool is destroyed. Kept as a diagnostic metric. amanvr@chromium.org woa-performance@google.com Records the number of times OnTransferSizeUpdated is called from different overriding methods. This UMA is useful to understand the distribution of calls being made to this method and will help see the reductions in IPC when we implement a prototype (b/257057692) where we remove unnecessary URLLoaderClient::OnTransferSizeUpdated IPCs. Disabled starting M137. ladan@chromium.org net-dev@chromium.org Records whether or not a url was upgraded by HSTS during an attempt to preconnect to the url and, if it was not upgraded, why. Recorded whenever GetHSTSRedirectForPreconnect() is called and kHstsTopLevelNavigationsOnly is disabled. ashleynewson@chromium.org src/android_webview/OWNERS Logged everytime an app uses reflection to invoke the ProxyChangeListener.ProxyReceiver BroadcastReceiver's onReceive method. This method exists for backwards compatibility with WebView apps that have not migrated to using newer proxy configuration APIs. This histogram is used to identify if it is still used in the app ecosystem or if it can be removed. Apps may trigger the method zero, one, or multiple times per session. Only logs to the false bucket. Only logged for Android WebView. barkinunal@google.com src/android_webview/OWNERS woa-performance@google.com Records the time taken to complete the registration of the ProxyChangeListener on Android. This measures the duration from when registration is initiated until it is complete. Note that there are two binder calls involved, as there are two receivers being registered. The metric covers the combined timing of these two calls. The combined timing won't tell whether one was slow or both were slow, but they are likely correlated. This metric is intended to help with judging priorities. This metric is only recorded on Android, for which a high-resolution clock is always available. dschinazi@chromium.org The time taken from requesting a PAC script to receiving the first byte of the response body on successful fetches. This does not include time spent doing proxy auto-discovery, or failed attempts at retrieving PAC scripts. If the response body was empty, this measures the total time to fetch the (empty) response body. nidhijaju@chromium.org blink-network-stack@google.com Records the number of active sessions that are marked as going away because {GoingAwayReason}. dschinazi@chromium.org src/net/quic/OWNERS The number of active QUIC sessions before we activate a new QUIC session. Warning: this histogram was expired from 09/12/2022 to 03/22/2023; data may be missing. dschinazi@chromium.org src/net/quic/OWNERS Tracks the number of QUIC alt-svc advertisements parsed by Chrome in Google format vs in IETF format. renjietang@chromium.org src/net/OWNERS Log the connection error code when QUIC stream handle closes. renjietang@chromium.org src/net/OWNERS Log the net error code when QUIC stream handle closes. renjietang@chromium.org src/net/OWNERS Log the stream error code when QUIC stream handle closes. bnc@chromium.org src/net/quic/OWNERS The delay in decoding received headers due to having to wait for dynamic table entries to arrive on the QPACK encoder stream. Zero if header block could be decoded as soon as it was received. This is recorded for connections made to {ServerType}. dschinazi@chromium.org src/net/OWNERS Records whether an HTTP Datagram was dropped when trying to send a UDP payload over the QUIC stream. Datagram is dropped if the QUIC session has not yet received SETTINGS_H3_DATAGRAM frame and therefore does not support sending or receiving datagrams. renjietang@chromium.org src/net/OWNERS Log success rate of processing trailing headers. renjietang@chromium.org src/net/quic/OWNERS The reason for the existing path validation when a new multi-port path creation is attempted but failed because of existing validation. renjietang@chromium.org src/net/quic/OWNERS The status of the multi-port path when the client initiates the migration. renjietang@chromium.org src/net/OWNERS When multi-port is enabled, log whether the server allows active migration that's required for multi-port wub@chromium.org src/net/quic/OWNERS The write result returned by the packet writer when the connection attempts to write a packet. dschinazi@chromium.org src/net/quic/OWNERS The number of active sessions that are degrading on network changes. The suffix specifies the network change. Emitted when there are more than one active sessions. The histogram is emitted when a network change happens which could affect the current default network, which includes the default network being changed, alternate networks being connected. It, however, doesn't emit when disconnecting an alternate network because that does not affect the current default network. {PlatformNotifications} dschinazi@chromium.org src/net/quic/OWNERS The number of active QUIC sessions when a network change is detected. The suffix specifies the network change. The histogram is emitted when a network change happens which could affect the current default network, which includes the default network being changed, alternate networks being connected. It, however, doesn't emit when disconnecting an alternate network because that does not affect the current default network. dschinazi@chromium.org src/net/quic/OWNERS The number of all sessions that detectes path degradation until a network change happens, including both active sessions and closed sessions. The suffix specifies the network change. The histogram is emitted when a network change happens which could affect the current default network, which includes the default network being changed, alternate networks being connected. It, however, doesn't emit when disconnecting an alternate network because that does not affect the current default network. {PlatformNotifications} dschinazi@chromium.org src/net/quic/OWNERS The total number of sessions that reports path degradation before a network change happens, including both still active sessions and closed sessions. The histogram is emitted when a network change happens which could affect the current default network, which includes the default network being changed, alternate networks being connected. It, however, doesn't emit when disconnecting an alternate network because that does not affect the current default network. dschinazi@chromium.org src/net/quic/OWNERS The number of total sessions that has been tracked since the first speculative error, including both active and closed sessions. Only recorded if non-zero. The histogram is emitted when a network change happens which could affect the current default network, which includes the default network being changed, alternate networks being connected. It, however, doesn't emit when disconnecting an alternate network because that does not affect the current default network. dschinazi@chromium.org src/net/quic/OWNERS The number of total sessions that has been closed due to QUIC_PACKET_WRITE_ERROR before first path degradation detection. The histogram is emitted when session path degrading is detected for the first time on the current network and there was a session closed with a speculative signal (QUIC_PACKET_WRITE_ERROR). dschinazi@chromium.org src/net/quic/OWNERS The percentage of active sessions that are degrading on network changes. The suffix specifies the network change. Emitted when there are at least two active sessions. The histogram is emitted when a network change happens which could affect the current default network, which includes the default network being changed, alternate networks being connected. It, however, doesn't emit when disconnecting an alternate network because that does not affect the current default network. {PlatformNotifications} dschinazi@chromium.org src/net/quic/OWNERS The percentage of all sessions that detectes path degradation until a network change happens, including both active sessions and closed sessions. Always emit 0 when there is no session reporting path degrading to speculate a connectivity failure on the current network. The suffix specifies the network change. The histogram is emitted when a network change happens which could affect the current default network, which includes the default network being changed, alternate networks being connected. It, however, doesn't emit when disconnecting an alternate network because that does not affect the current default network. {PlatformNotifications} dschinazi@chromium.org src/net/quic/OWNERS When a packet write error is detected by the session, if path degradation has been detected already on the same session. The histogram is emitted when a session encounters a QUIC_PACKET_WRITE_ERROR and is soon to be closed. dschinazi@chromium.org src/net/quic/OWNERS Tracks failure reasons to read/load/write of QuicServerInfo to and from disk cache. {NetQuicDiskCacheBackend} dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS The reason a QUIC packet could not be processed, logged for each dropped packet. WARNING: this histogram was expired from 2025-01-05 to 2026-06-24; data may be missing. dschinazi@chromium.org src/net/quic/OWNERS The number of QUIC packets received by a QUIC connection whose handshake was not confirmed when that connection is closed. Warning: this histogram was expired from 09/12/2022 to 03/22/2023; data may be missing. dschinazi@chromium.org src/net/quic/OWNERS The state of a QUIC connection's crypto hanshake as it progresses from starting to confirmation or failure. renjietang@chromium.org src/net/quic/OWNERS Log the success rate of HTTP layer reading QUIC response headers, for investigating b/177893733. bnc@chromium.org src/net/quic/OWNERS Stream error of QuicSpdyStream when QuicHttpStream::ComputeResponseStatus() returns ERR_QUIC_PROTOCOL_ERROR, for investigating b/177893733. renjietang@chromium.org src/net/quic/OWNERS The percentage comparison between alternative port probe failure when default path is degrading and in general. renjietang@chromium.org src/net/quic/OWNERS The percentage comparison between alternative port response time when default path is degrading and response time in general. dmcardle@chromium.org renjietang@chromium.org src/net/quic/OWNERS The number of alternative port probing failure when the default path is degrading. dmcardle@chromium.org renjietang@chromium.org src/net/quic/OWNERS The number of alternative port probing failure when the default path is healthy. dmcardle@chromium.org renjietang@chromium.org src/net/quic/OWNERS The number of alternative port probing attempts. dmcardle@chromium.org renjietang@chromium.org src/net/quic/OWNERS The number of successful alternative port probes. dschinazi@chromium.org src/net/quic/OWNERS Time from when QUIC detects performance degradation on the current network to when QUIC notices a new network is connected. dschinazi@chromium.org src/net/quic/OWNERS Time from when QUIC detects performance degradation on the current network to when QUIC notices the current network is disconnected. dschinazi@chromium.org src/net/quic/OWNERS The time from QUIC detecting network performance degradation to an alternative network becoming default. dschinazi@chromium.org src/net/quic/OWNERS The time from network disconnection to an alternative network becoming default. dschinazi@chromium.org src/net/quic/OWNERS The number of times ServerConfigUpdateMessages ignored. dschinazi@chromium.org src/net/quic/OWNERS Records whether queue is at maximum capacity when an HTTP/3 datagram arrives on the QUIC stream in client socket. A QUIC connection parses the Datagram frame and passes it to the QUIC SPDY session, which then extracts the quarter stream ID to determine which stream to invoke `OnHttp3Datagram` method on. The incoming datagram is dropped if full and added to queue otherwise. dschinazi@chromium.org src/net/quic/OWNERS The state of the QUIC server information when it's loaded from the disk cache. dschinazi@chromium.org src/net/quic/OWNERS The number of pending stream requests aborted when the session is closed. Pending streams are those which could not be created immediately because there were too many active streams. bnc@chromium.org src/net/OWNERS Records whether the QUIC session received an Accept-CH entry for the given origin, which GetAcceptChViaAlps() returns. Logged every time QuicHttpStream::GetAcceptChViaAlps() is called, typically once for each request that is served over HTTP/3. bnc@chromium.org src/net/OWNERS Records both whether there are any valid origin fields and whether there are any invalid ones in the ACCEPT_CH frame. Logged during the TLS handshake of a QUIC connection if ALPS is negotiated, once for every ACCEPT_CH HTTP/3 frame received (possibly zero, typically not more than one). renjietang@chromium.org src/net/quic/OWNERS The number of active streams when the quic session decides to go away on path degrading. bashi@chromium.org blink-network-stack@google.com The number of connection attempts a QuicSessionPool::AsyncDnsJob started before it settled. Recorded once per job, when {JobOutcome}. A job recorded with zero attempts failed before it could start a connection attempt or succeeded by using an existing session. Only recorded when the AsyncDnsQuicJob feature is enabled. bashi@chromium.org blink-network-stack@google.com Time since the successful connection attempt started, measured at {MeasurementPoint}. The FinalDnsResult variant records zero if the attempt started after DNS finished. The JobSuccessWithDnsInFlight variant is a lower bound on the time until the final DNS result because job success cancels DNS. Recorded once per successful job that used a connection attempt. Only recorded when the AsyncDnsQuicJob feature is enabled. bashi@chromium.org blink-network-stack@google.com How a successful QuicSessionPool::AsyncDnsJob obtained its session. The source is an active session for the same key, IP pooling, the first attempt by the initial connector, a later attempt by the initial connector, or any attempt by the connector created when the slow timer fired. Recorded once per successful job. Only recorded when the AsyncDnsQuicJob feature is enabled. bashi@chromium.org blink-network-stack@google.com The time from the start of a QuicSessionPool::AsyncDnsJob's first connection attempt to the job's failure. Recorded once per failed job that started at least one attempt. Jobs that fail before starting an attempt are not recorded. Only recorded when the AsyncDnsQuicJob feature is enabled. dschinazi@chromium.org src/net/quic/OWNERS True if the result of reading a packet from the network was ERR_IO_PENDING. Recorded for each packet when Read() returns. Warning: this histogram was expired from 09/12/2022 to 03/22/2023; data may be missing. dschinazi@chromium.org src/net/quic/OWNERS The number of BLOCKED frames recevied by a QuicSession when the session is closed. dschinazi@chromium.org src/net/quic/OWNERS The number of BLOCKED frames sent by a QuicSession when the session is closed. dschinazi@chromium.org src/net/quic/OWNERS The network error code returned to a QUIC session from the CertVerifier when a certification verification is attempted. dschinazi@chromium.org src/net/quic/OWNERS Time spent verifying a certificate when racing cert veriifcation with host resolution. dschinazi@chromium.org src/net/quic/OWNERS The final maximum packet size of the connection used by the client before the session is closed. WARNING: this histogram was expired from 2025-04-10 to 2026-06-24; data may be missing. dschinazi@chromium.org src/net/quic/OWNERS The network error code which resulted in all sessions being closed. wub@chromium.org src/net/quic/OWNERS For a QUIC session that is closed by a public reset, where the handshake was {HandshakeType}, this boolean shows whether it is a public reset sent from a Google server or not. Warning: this histogram was expired between 2024-06-02 and 2026-08-07 so some data is missing. wub@chromium.org fayang@chromium.org src/net/quic/OWNERS For a QUIC session that is closed by QUIC_TOO_MANY_RTOS at client, this is the number of packets it received before it is closed. wub@chromium.org fayang@chromium.org src/net/quic/OWNERS For a QUIC session that is closed by QUIC_TOO_MANY_RTOS at client, this is the number of packets it sent before it is closed. dschinazi@chromium.org src/net/quic/OWNERS True if the QUIC session is closed during the call to InitializeSession, logged for each session just after InitializeSession is called. dschinazi@chromium.org src/net/quic/OWNERS The network error code which resulted in the session being closed. renjietang@chromium.org src/net/quic/OWNERS When a new stream frame is added to a QUIC packet, record whether it's coalesced with the existing stream frames. dschinazi@chromium.org src/net/quic/OWNERS True if a QUIC connection connected successfully after having been broken. Only emitted after a QUIC connection attempt to a server that was previous marked as broken. dschinazi@chromium.org src/net/quic/OWNERS The QUIC error which caused a QUIC connection to be closed before the hanshake was confirmed, in the case where no packets were received. This provides a breakdown of the entires in Net.QuicSession.ConnectionClose.HandshakeNotConfirmed.Reason where the value is BLACK_HOLE. dschinazi@chromium.org src/net/quic/OWNERS The QUIC error which caused a QUIC connection to be closed before the hanshake was confirmed, in the case where at least 1 packet was received. This provides a breakdown of the entires in Net.QuicSession.ConnectionClose.HandshakeNotConfirmed.Reason where the value is UNKNOWN. dschinazi@chromium.org src/net/quic/OWNERS The reason a QUIC handshake failed. dschinazi@chromium.org src/net/quic/OWNERS The number of streams open when a QUIC session crypto handshake timed out. Warning: this histogram was expired from 09/12/2022 to 03/22/2023; data may be missing. dschinazi@chromium.org src/net/quic/OWNERS The number of streams open when a QUIC session timed out. dschinazi@chromium.org src/net/quic/OWNERS The number of total streams created when a QUIC session crypto handshake timed out. dschinazi@chromium.org src/net/quic/OWNERS The QUIC application transport error code in a CONNECTION_CLOSE frame received from the server{GQuicMissing}. This is recorded for connections made to {ServerType}{HandshakeType}. davidben@chromium.org dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS The QUIC transport error code in a CONNECTION_CLOSE frame received from the server{GQuicMissing}. This is recorded for connections made to {ServerType}{HandshakeType}. Warning: the ConnectionCloseErrorCodeServerIetfTransportGoogle variant was expired from 09/12/2022 to 03/22/2023; data may be missing. davidben@chromium.org dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS The QUIC error code which resulted in the QUIC connection being closed by the {Closer}. This is recorded for connections made to {ServerType}{HandshakeType}. davidben@chromium.org dschinazi@chromium.org src/net/quic/OWNERS fayang@chromium.org src/net/quic/OWNERS Duration of connections received PUBLIC_RESET. renjietang@chromium.org suzukikeita@chromium.org src/net/quic/OWNERS The duration of a QUIC connection starting from 1-RTT key available to connection close for {HostType} with {Initiator}. This is recorded when the QUIC connection is closed. dschinazi@chromium.org src/net/quic/OWNERS Whether QUIC session's connection is flow control blocked when a PING message is sent to server. nidhijaju@chromium.org src/net/quic/OWNERS True if an active session already exists for a given Ip during connection. Recorded when attempting to create a new session. Warning: this histogram was expired from 2024-02-04 to 2024-03-28; data may be missing. dschinazi@chromium.org src/net/quic/OWNERS The result of a QUIC connection migration attempt. The suffix specifies the cause of connection migration. {ConnectionMigrationCause} dschinazi@chromium.org src/net/quic/OWNERS The IP Address family of this connection, as reported by the server. dschinazi@chromium.org src/net/quic/OWNERS Samples of the number of round-trips needed by a QUIC connection before a request could be sent by the client. {QuicPortSelection} dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS Count of errors during attempts to create a QUIC session (before even using the session). fayang@chromium.org src/net/quic/OWNERS Times of retransmission time out of connections with successful handshake. fayang@chromium.org src/net/quic/OWNERS Times of crypto retransmission timeout of connections with failed handshake. renjietang@chromium.org src/net/quic/OWNERS The number of draining streams when the quic session decides to go away on path degrading. dschinazi@chromium.org src/net/quic/OWNERS The number of duplicate packets recevied by a QuicSession when the session is closed. martinduke@google.com rch@chromium.org The combination of ECN marks sent on a single QuicSession. Recorded when the session is closed. fayang@chromium.org src/net/quic/OWNERS The elapsed time between starting the crypto handshake and installing the first encryption key that can carry application data (either 0-RTT or 1-RTT). dschinazi@chromium.org src/net/quic/OWNERS The value of the TCP cubic sender's CWND when the session is closed. dschinazi@chromium.org src/net/quic/OWNERS Logged every time we parse a QUIC Finch config, and logs whether the config was valid. Invalid configs are most likely to be obsolete configs. Note that this is still logged (as true) when no variations config is present. Warning: this histogram was expired from 09/12/2022 to 03/22/2023; data may be missing. dschinazi@chromium.org src/net/quic/OWNERS Triggered any time Finch configures an obsolete QUIC version. Only recorded when obsolete versions are not allowed (which is the default). fayang@chromium.org src/net/quic/OWNERS Result of finding matching IP session when DNS resolution finishes. dschinazi@chromium.org src/net/quic/OWNERS This measures the time gap between the most recent write error and the network disconnection. dschinazi@chromium.org src/net/quic/OWNERS Whether QuicGoAwayFrame is received from server for connection migration due to client's port change. Logged when a QuicGoAwayFrame with error code QUIC_ERROR_MIGRATING_PORT is received. suzukikeita@chromium.org blink-network-stack@google.com The session establishment reason for a QUIC session to GoogleSearch. This is recorded when closing a session to GoogleSearch, indicating if it was {Used}{Initiator}. suzukikeita@chromium.org blink-network-stack@google.com The session non-reuse reason for a QUIC session to GoogleSearch. This is recorded when closing the session to GoogleSearch, and when {Used} with {Initiator}. dschinazi@chromium.org src/net/quic/OWNERS The elapsed time between starting the crypto handshake, and receiving confirmation from the server. Warning: this histogram was expired from 2023-01-15 to 2023-11-08; data may be missing. davidben@chromium.org src/net/quic/OWNERS The elapsed time between starting the crypto handshake, and receiving confirmation from the server, measured only for ECH-capable servers. davidben@chromium.org hchao@chromium.org src/net/quic/OWNERS The elapsed time between starting the crypto handshake, and receiving confirmation from the server, measured only for servers responding with the Server Padding extension. estark@chromium.org davidben@chromium.org src/net/quic/OWNERS The elapsed time between starting the crypto handshake, and receiving confirmation from the server, measured only for servers that advertise TLS Trust Anchor IDs in DNS, {IsResumption}. Warning: this histogram was expired from 2025-10-26 to 2026-02-12; data may be missing. dschinazi@chromium.org src/net/quic/OWNERS The handshake status when the session attempts to migrate the connection according to different migration causes. {ConnectionMigrationCause} fayang@chromium.org src/net/quic/OWNERS The handshake status when the session attempts to validate received server preferred address. dschinazi@chromium.org src/net/quic/OWNERS The handshake status when the session attempts to port migration on path degrading. dschinazi@chromium.org src/net/quic/OWNERS Whether path degrading has been detected by the time a QUIC connection is closed due to handshake timeout. dschinazi@chromium.org src/net/quic/OWNERS Header compression ratio as percentage for received headers using HPACK. Warning: this histogram was expired from 09/12/2022 to 03/22/2023; data may be missing. dschinazi@chromium.org src/net/quic/OWNERS Header compression ratio as percentage for sent headers using HPACK. Warning: this histogram was expired from 09/12/2022 to 03/22/2023; data may be missing. dschinazi@chromium.org src/net/quic/OWNERS Header compression ratio as percentage for received headers using QPACK. Warning: this histogram was expired from 09/12/2022 to 03/22/2023; data may be missing. dschinazi@chromium.org src/net/quic/OWNERS Header compression ratio as percentage for sent headers using QPACK. dschinazi@chromium.org src/net/quic/OWNERS The elapsed time between the end of resolving the DNS name of the server for a QUIC connection, and receiving crypto handshake confirmation from the server. Logged after crypto handshake is confirmed. WARNING: this histogram was expired from 2025-03-09 to 2026-06-24; data may be missing. dschinazi@chromium.org src/net/quic/OWNERS The number packets recevied by a QuicSession with an incorrect connection id when the sesesion is closed. dschinazi@chromium.org src/net/quic/OWNERS The source for the initial RTT estimate recorded when a QUIC session is created. dschinazi@chromium.org src/net/quic/OWNERS The number of key updates processed per connection, recorded on connection close. Only logged for connections which reached the 1-RTT level and supported key update. dschinazi@chromium.org src/net/quic/OWNERS Count of packets received that appeared to attempt a key update but failed decryption which have been received since the last successfully decrypted packet, recorded on connection close. Only logged for connections which reached the 1-RTT level and supported key update. mattm@chromium.org src/net/quic/OWNERS The reason why each key update occured. mattm@chromium.org src/net/quic/OWNERS Records the success or failure of the last key update for connections where the key update was initiated by the {Group} side. Failure is defined as having sent at least one packet in the new key phase but never receiving an ack for any packet sent in the new key phase before the connection was closed. fayang@chromium.org src/net/quic/OWNERS Time from handshake completion to last packet was sent for connections received PUBLIC_RESET. fayang@chromium.org src/net/quic/OWNERS A bitfield representing the last sent packet content before connection receives PUBLIC_RESET. fayang@chromium.org src/net/quic/OWNERS Max consecutive retransmission timeout before making forward progress. fayang@chromium.org src/net/quic/OWNERS Max consecutive retransmission timeout before making forward progress when blackhole has been detected. dschinazi@chromium.org src/net/quic/OWNERS The maximum packet sequence number reordering observed by a QUIC connection. dschinazi@chromium.org src/net/quic/OWNERS The ratio of the maximum reordering time of a QUIC packet to the min rtt. dschinazi@chromium.org src/net/quic/OWNERS The ratio of the maximum reordering time of a QUIC packet to the min rtt, only for those sessions with a min rtt larger than 100 ms. renjietang@chromium.org src/net/quic/OWNERS Records whether the migration of the QuicSession to a new socket is successful. dschinazi@chromium.org src/net/quic/OWNERS The minimum RTT observed during the life of a QUIC connection. dschinazi@chromium.org src/net/quic/OWNERS The number of MTU probes sent by the client during the session. Logged when the connection is destroyed. WARNING: this histogram was expired from 2025-04-10 to 2026-06-24; data may be missing. renjietang@chromium.org suzukikeita@chromium.org src/net/quic/OWNERS The number of active streams when idle timeout is detected and connection is being closed for {HostType}. This is recorded when the QUIC connection is closed because of idle timeout. This metric has been re-enabled on M138, and any data prior to this is missing. This histogram has been expired between 2025-08-24 and 2026-07-01, thus does not contain data for this period. renjietang@chromium.org src/net/quic/OWNERS The number of times that a session enters path-degrading state on the default path. renjietang@chromium.org src/net/quic/OWNERS The number of times that a session makes forward progress after the original default path is degrading. renjietang@chromium.org src/net/quic/OWNERS The number of successful migrations for a QUIC session. dschinazi@chromium.org src/net/quic/OWNERS Number of connection migrations exercised before the connection receives a public reset. dschinazi@chromium.org src/net/quic/OWNERS The number of QUIC streams opened when a new QUIC stream is created. dschinazi@chromium.org src/net/quic/OWNERS The number of pending stream requests when a new stream request is added to the pending list. Pending streams are those which could not be created immediately because there were too many active streams. renjietang@chromium.org suzukikeita@chromium.org src/net/quic/OWNERS The number of pings a QUIC connection sends during its lifetime. This is recorded when the QUIC connection is closed for {HostType}. This metric has been re-enabled on M138, and any data prior to this is missing. fayang@chromium.org src/net/quic/OWNERS The number of origins in received ORIGIN frames. This is recorded for each received ORIGIN frame. If multiple ORIGIN frames are received over the duration of a connection, this count is recorded for each frame. renjietang@chromium.org suzukikeita@chromium.org src/net/quic/OWNERS The number of streams waiting to write when idle timeout is detected and connection is being closed for {HostType}. This is recorded when the QUIC connection is closed because of idle timeout. This metric has been re-enabled on M138, and any data prior to this is missing. dschinazi@chromium.org src/net/quic/OWNERS The total number of streams created by the client when the session is closed. fayang@chromium.org src/net/quic/OWNERS The result of a QUIC server preferred address migration attempt. dschinazi@chromium.org src/net/quic/OWNERS The number of missing packets between the current received packet and the previously largest received packet sequence number, when the current received packet had a lower sequence number than the previously received packet sequence number. dschinazi@chromium.org src/net/quic/OWNERS The number of times the current received packet had a lower sequence number than the previously received packet sequence number, and the size of the current packet is larger than the size of the previous packet. dschinazi@chromium.org src/net/quic/OWNERS The number of times the current received packet had a lower sequence number than the previously received packet sequence number. dschinazi@chromium.org src/net/quic/OWNERS The number of missing packets between the current received packet and the previously largest received packet sequence number. dschinazi@chromium.org src/net/quic/OWNERS The number of missing packets between the last received packet before a PING frame sent and the first packet received after PING sent. dschinazi@chromium.org src/net/quic/OWNERS The number of missing packets between the current received packet and the previously largest received packet sequence number, as reported by the remote end of the connection. dschinazi@chromium.org src/net/quic/OWNERS The ratio of the number of missing packets, to the maximum packet sequence number received, for QUIC connections longer than 21 packets received via Warning: this histogram was expired from 2023-10-01 to 2024-03-06; data may be missing. {Net_QuicSession_PacketReceived_CONNECTION_TYPE} dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS dschinazi@chromium.org src/net/quic/OWNERS The number of packets retransmitted per 1000. Only sessions with packets_sent >= 100 are included. Recorded in session destructor. dschinazi@chromium.org src/net/quic/OWNERS The time taken to Write() a QUIC packet to the socket. Recorded for each packet when it is sent. The suffix specifies if the write completed synchonously or asynchronously. Warning: this histogram was expired from 2023-11-19 to 2024-03-06 and from 2024-09-06 to 2026-03-26; data may be missing. {IOMode} martinduke@google.com src/net/quic/OWNERS The result of path validation according to different migration causes. {ConnectionMigrationCause} bashi@chromium.org src/net/quic/OWNERS The wait time of the pending stream when it is finally serviced. Pending streams are those which could not be created immediately because there were too many active streams. Recorded when a QUIC session is ready open a new stream on top of the session. Warning: this histogram was expired from 2023-02-12 to 2026-03-09; data may be missing. dschinazi@chromium.org src/net/quic/OWNERS The platform notification received by QUIC when network change happens. dschinazi@chromium.org src/net/quic/OWNERS The result of a QUIC port migration attempt. dschinazi@chromium.org src/net/quic/OWNERS True if the QUIC session prefers to use AES GCM because of hardware support. dschinazi@chromium.org src/net/quic/OWNERS Number of retries to send connectivity probing packet on new path until success. dschinazi@chromium.org src/net/quic/OWNERS The wait time to receive a matched probing response to mark connectivity probing as successful. dschinazi@chromium.org src/net/quic/OWNERS The ordinality of a header list within a connection during the encoding of which unacknowledged dynamic table entries could not be referenced due to the limit on the number of blocked streams. dschinazi@chromium.org src/net/quic/OWNERS The ordinality of a header list within a connection during the encoding of which at least one dynamic table insertion was blocked. dschinazi@chromium.org src/net/quic/OWNERS The ordinality of a header list within a connection during the encoding of which no dynamic table insertion was blocked. dschinazi@chromium.org src/net/quic/OWNERS The ordinality of a header list within a connection during the encoding of which the limit on the number of blocked streams did not prevent referencing unacknowledged dynamic table entries. dschinazi@chromium.org src/net/quic/OWNERS Version of the QUIC protocol used for this connection. dschinazi@chromium.org src/net/quic/OWNERS The network error code returned when attempting to read on the current network from a QUIC connection after handshake has been confirmed. dschinazi@chromium.org src/net/quic/OWNERS The network error code returned when attempting to read from a QUIC connection. The suffix specifies the condition/source of the read error. {ReadErrorSourceNetwork} aymanm@google.com ricea@chromium.org The number of UDP packets successfully read in a single batch read (`ReadMultiple` / `recvmmsg`) operation on a QUIC UDP socket. Recorded each time `QuicChromiumPacketReader` processes a successful multi-packet read result when the `QuicUseReadMultiple` feature is enabled. Note: Because packet reading occurs at high frequency, this metric is subsampled and recorded for 10% of successful read operations. dschinazi@chromium.org src/net/quic/OWNERS The value of the SETTINGS_QPACK_BLOCKED_STREAMS parameter received on an HTTP/3 connection, if any. dschinazi@chromium.org src/net/quic/OWNERS The number of SETTINGS parameters received on an HTTP/3 connection, incremented by one (the SETTINGS frame can be empty). dschinazi@chromium.org src/net/quic/OWNERS The value of the SETTINGS_MAX_HEADER_LIST_SIZE parameter received on an HTTP/3 connection, if any. dschinazi@chromium.org src/net/quic/OWNERS The value of the SETTINGS_QPACK_MAX_TABLE_CAPACITY parameter received on an HTTP/3 connection, if any. Warning: this histogram was expired from 09/12/2022 to 03/22/2023; data may be missing. dschinazi@chromium.org src/net/quic/OWNERS The number of reserved SETTINGS parameters (that is, with identifier in the form 0x1f * N + 0x21 for a non-negative integer value of N) received on an HTTP/3 connection, incremented by one (there might be zero reserved identifiers). Defined at https://quicwg.org/base-drafts/draft-ietf-quic-http.html#name-defined-settings-parameters. These should not be treated specially on the receive side, because they are sent to exercise the requirement that unknown identifiers are ignored. Here an exception is made for logging only, to understand what kind of identifiers are received. dschinazi@chromium.org src/net/quic/OWNERS True if the QUIC REJ message received from the server contains a proof. dschinazi@chromium.org src/net/quic/OWNERS The length in bytes of a QUIC REJ message received from the server. yfriedman@chromium.org src/net/quic/OWNERS Records whether a request to RequestSession was blocked because the app was pre-frozen. Logged for every RequestSession call on Android. thesalsa@google.com src/net/quic/OWNERS The age of the session ticket when the client attempted 0-RTT and the handshake completed. Recorded for 0-RTT attempts that were {Status}. The maximum valid age of session tickets is capped at 7 days by the TLS 1.3 spec, though in practice this may be set lower by the client. Capped to 14 days in this histogram. https://datatracker.ietf.org/doc/html/rfc8446#section-4.6.1 dschinazi@chromium.org src/net/quic/OWNERS The number of consecutive times a packet was retried after a write error, using exponential backoff. dschinazi@chromium.org src/net/quic/OWNERS The QUIC error code which resulted in a stream being reset by the client. dschinazi@chromium.org src/net/quic/OWNERS The QUIC error code which resulted in a stream being reset by the server. dschinazi@chromium.org src/net/quic/OWNERS The number of request for secure resources over QUIC sessions. True if the session is secure, false if it is not. dschinazi@chromium.org src/net/quic/OWNERS When a SHLO handshake message is received, whether the client IP address and port number in it differ from the local client IP address and port number. In the comparison, the first address is the one in SHLO and the second address is the client IP address. dschinazi@chromium.org src/net/quic/OWNERS The size in bytes of an encrypted QUIC packet sent to the peer at the specified encryption level. {EncryptionLevel} dschinazi@chromium.org src/net/quic/OWNERS The largest packet which the client received from the server during the session. Warning: this histogram was expired from 2023-11-19 to 2024-04-11; data may be missing. WARNING: this histogram was expired from 2025-04-10 to 2026-06-24; data may be missing. suzukikeita@chromium.org blink-network-stack@google.com Records which combination of fields in `QuicSessionKey` mismatched with the key of existing session or an in-flight session attempt to {HostType} with the same `ServerId`. We only look at the combinations of frequent fields which cause mismatches, retrieved by `MismatchedField` histograms. Recorded every time when we attempt to create a QUIC session. suzukikeita@chromium.org blink-network-stack@google.com Records which fields in `QuicSessionKey` mismatched with the key of existing session or an in-flight session attempt to {HostType} with the same `ServerId`. Recorded every time when we attempt to create a QUIC session. suzukikeita@chromium.org blink-network-stack@google.com Records the total number fields in `QuicSessionKey` which mismatched with the key of existing session or an in-flight session attempt to {HostType} with the same `ServerId`. Recorded every time when we attempt to create a QUIC session. suzukikeita@chromium.org blink-network-stack@google.com Records whether `QuicSessionKey` had an existing session or an in-flight session attempt to {HostType}. Recorded every time when we attempt to create a QUIC session. dschinazi@chromium.org src/net/quic/OWNERS The final smoothed RTT observed during the life of a QUIC connection. renjietang@chromium.org src/net/quic/OWNERS True if fresh host lists contains stale host used in racing connection. renjietang@chromium.org src/net/quic/OWNERS The time a QUIC connection from invalid stale dns result takes to do Crypto handshake. This starts on DoConnect and ends when the connection finishes. renjietang@chromium.org src/net/quic/OWNERS Records when a STATELESS_RESET is received after a connectivity probing is cancelled. renjietang@chromium.org src/net/quic/OWNERS When probing is cancelled, records whether a STATELESS_RESET is received. dschinazi@chromium.org src/net/quic/OWNERS The QUIC error code that we sent in a STOP_SENDING frame. dschinazi@chromium.org src/net/quic/OWNERS The QUIC error code that we received in a STOP_SENDING frame. dschinazi@chromium.org src/net/quic/OWNERS The QUIC error code which resulted in the stream (and connection) being closed by the client after the handshake was confirmed. dschinazi@chromium.org src/net/quic/OWNERS The QUIC error code which resulted in the stream (and connection) being closed by the client after the handshake was confirmed. dschinazi@chromium.org src/net/quic/OWNERS Whether any of QUIC session's streams are flow control blocked when a PING message is sent to server. dschinazi@chromium.org src/net/quic/OWNERS The number of stream frames received which were duplicates, out of every 1000 stream frames received. Only for QUIC sessions which received at least 100 packets. dschinazi@chromium.org src/net/quic/OWNERS The percentage of stream frames received which were duplicates. Only for QUIC sessions which received at least 100 packets. dschinazi@chromium.org src/net/quic/OWNERS The percentage of stream frames received which were duplicates. Only for QUIC sessions which received fewer than 100 packets. dschinazi@chromium.org src/net/quic/OWNERS The number of stream frames received which were duplicates, out of every 1000 stream frames received. Only for QUIC sessions which received fewer than 100 packets. dschinazi@chromium.org src/net/quic/OWNERS If a QUIC connection timed out with open streams, this contains a count of consecutive PTOs. dschinazi@chromium.org src/net/quic/OWNERS If a QUIC connection timed out with open streams, this contains a count of consecutive RTOs. dschinazi@chromium.org src/net/quic/OWNERS If a QUIC connection timed out with open streams, this contains a count of consecutive TLPs. dschinazi@chromium.org src/net/quic/OWNERS If a QUIC connection timed out with open streams, this will be true when the connection has unacked packets. dschinazi@chromium.org src/net/quic/OWNERS If a QUIC connection timed out locally with open streams, this contains the local port number for the connection. renjietang@chromium.org src/net/quic/OWNERS The time a QuicSessionPool Job takes from host rersolution to connection confirmation. dschinazi@chromium.org src/net/quic/OWNERS True if more than 100 streams are open when a new stream is activated. dschinazi@chromium.org src/net/quic/OWNERS When a QUIC packet is sent at initial encryption that is smaller than the minimum size, this records the size difference (minimum_size - observed_size). dschinazi@chromium.org src/net/quic/OWNERS The number of undecryptable packets recevied by a QuicSession when the session is closed. Includes packets where a decryption key was not present. dschinazi@chromium.org src/net/quic/OWNERS The number of undecryptable packets recevied by a QuicSession when the session is closed. Only includes packets where a decryption key was present for the packet's encryption level. renjietang@chromium.org src/net/quic/OWNERS The time a QUIC connection that will be confirmed takes for Crypto handshake. The event starts on DoConnect, and ends when the connection finishes. dschinazi@chromium.org src/net/quic/OWNERS Time spent verifying the signature and certificate chain. This is logged whenever QUIC verifies the certificate chain and signature during crypto handshake. Warning: this histogram was expired from 09/12/2022 to 03/22/2023; data may be missing. dschinazi@chromium.org src/net/quic/OWNERS The network error code returned when attempting to write to a QUIC connection. WARNING: this histogram was expired from 2025-04-10 to 2026-06-24; data may be missing. dschinazi@chromium.org src/net/quic/OWNERS The network error code returned when attempting to write to a QUIC connection after the handshake has been confirmed. dschinazi@chromium.org src/net/quic/OWNERS The most recent network error code when attempting to write to a QUIC connection when the network gets disconnected. dschinazi@chromium.org src/net/quic/OWNERS The reason QUIC packet writer needed to allocate a new buffer, rather than the expected case of reusing its existing one. dschinazi@chromium.org src/net/quic/OWNERS Indicates whether a QUIC handshake attempted or used 0-RTT on the connection, and if 0-RTT was not used, it provides reasons why. Records connections to {HostType} when a QUIC handshake is completed. bashi@chromium.org src/net/quic/OWNERS Whether 0-RTT was successfully used in the connection. Records connections to {HostType} when a QUIC handshake is completed. Warning: this histogram was expired from 2024-05-19 to 2024-10-25; data may be missing. bashi@chromium.org blink-network-stack@google.com Record the time that a QuicSessionAttempt takes to invoke the completion callback. Recorded when a QuicSessionAttempt called the completion callback. bashi@chromium.org blink-network-stack@google.com Record the time that a QuicSessionAttempt takes to create a session. Recorded when a QuicSessionAttempt created a session asynchronously. bashi@chromium.org blink-network-stack@google.com Record the time that a QuicSessionAttempt takes to perform a crypto handshake. Recorded when a QuicSessionAttempt performed a crypto handshake asynchronously. dschinazi@chromium.org src/net/quic/OWNERS Records if a new connection on the alternate network is attempted when a connection fails before handshake is confirmed. dschinazi@chromium.org src/net/quic/OWNERS Connection type of the default network when the connection fails with handshake not confirmed. dschinazi@chromium.org src/net/quic/OWNERS A count of how often a user hits 1, 2, 3, 4 and 5 bad packet loss events for port 443. Recorded whenever QUIC has bad packet loss (high packet loss on multiple consecutive connections). dschinazi@chromium.org src/net/quic/OWNERS Status of the connection on the non-default network when it finishes. A connection may stay on the non-default network during the initial setup when the platform's default network changes. dschinazi@chromium.org src/net/quic/OWNERS Records, for initial socket created for each new QUIC connection, whether the bound network matched the default network tracked by QuicSessionPool. dschinazi@chromium.org src/net/quic/OWNERS Rrecords reasons QUIC is disabled (for all ports), if sufficent recent connections experience: public reset post crypto handshake, or timeouts with streams open. QUIC is disabled until the next reboot of Chrome. Logged during session close. renjietang@chromium.org src/net/quic/OWNERS Log the location in QuicSessionPool where QUIC_PROTOCOL_ERROR is reported. dschinazi@chromium.org src/net/quic/OWNERS Connection type of the alternate network when a connection is about to be attempted when the original connection fails on the default network with handshake not confirmed. dschinazi@chromium.org src/net/quic/OWNERS Status of the connection which is created for migration before handshake is confirmed. dschinazi@chromium.org src/net/quic/OWNERS The network error code returned when the retried connection on the alternate network before handshake fails. dschinazi@chromium.org src/net/quic/OWNERS Captures whether the default network has been changed when the new connection succeeds on the "alternate" network. dschinazi@chromium.org src/net/quic/OWNERS Captures the maximum number of public resets post handshake that occurred within a window of recent connections (default 20). Will help inform the choice of threshold to disable QUIC for clients that experience pathalogical errors. Logged during session close. dschinazi@chromium.org src/net/quic/OWNERS Recorded whenever QUIC is disabled for a port due to repeated lossy connections (high packet loss on multiple consecutive connections). QUIC is disabled until next reboot of Chrome. dschinazi@chromium.org src/net/quic/OWNERS Captures the maximum number of connection timeouts with streams open that occurred within a window of recent connections (default 20). Will help inform the choice of threshold to disable QUIC for clients that experience pathalogical errors. Logged during session close. jkarlin@chromium.org src/net/base/registry_controlled_domains/OWNERS Whether a call to GetDomainAndRegistry was served from the cache or not. Recorded on a sampled (0.00001) subset of calls to GetDomainAndRegistry. ricea@chromium.org patilmayur@microsoft.com src/net/reporting/OWNERS This metric is recorded each time a report upload attempt is made, when the reports are serialized to JSON. It reflects the size of the report list after filtering based on the report body size. Reports with bodies exceeding kMaxReportBodySizeKB are excluded when the kExcludeLargeBodyReports feature is enabled. rodneyding@google.com src/net/reporting/OWNERS The type of header received for configuring reporting API endpoints, recorded when reporting API headers are successfully parsed. Warning: this histogram was expired from 2023-02-18 to 2024-04-09; data may be missing. ricea@chromium.org patilmayur@microsoft.com src/net/reporting/OWNERS Records the estimated memory usage, in kb, of the report body for each report being considered for upload to the Reporting API. This metric is recorded for all reports, before any filtering based on body size. ricea@chromium.org patilmayur@microsoft.com src/net/reporting/OWNERS Counts the total number of reports initially considered for serialization and upload to the Reporting API. This metric reflects the size of the report list before any filtering based on size of report body. ricea@chromium.org patilmayur@microsoft.com src/net/reporting/OWNERS Records the estimated memory usage, in kb, of the complete serialized JSON payload containing the reports being prepared for upload to the Reporting API. This metric reflects the size of the JSON after filtering based on report body size. rodneyding@google.com src/net/reporting/OWNERS The type of header that produced the endpoint used for uploading reports for the Reporting API, recorded when reports have been succesfully uploaded, counted once for each report. tov@google.com chrome-privacy-core@google.com The length of the requested URL, excluding the fragment. Only logged for requests where the URL length exceeds 8192 characters. Recorded when the network loader (URLLoader) request is initiated in the Network Service. olivierli@chromium.org woa-performance@google.com Counts the number of RestrictedCookieManager::GetCookiesString IPCs received over 30 second periods at the Network Service. We start recording the metric when the Network Service is created and keep updating it for the entire lifetime of the Network Service process. victorvianna@google.com woa-performance-team+alerts@google.com Records the time spent executing RestrictedCookieManager::GetCookiesString, plus its callback if the callback is invoked synchronously. Subsampled to 1/1000. See kHistogramSampleProbability in net/cookies/cookie_constants.h for details. Not reported for clients with low-resolution clocks. dylancutler@google.com src/net/cookies/OWNERS Logs the number of partitioned cookies accessed by a script. Logged only when partitioned cookies are enabled. Subsampled to 1/1000. See kHistogramSampleProbability in net/cookies/cookie_constants.h for details. Warning: this histogram was expired from 2026-07-22 to 2026-08-25; data may be missing. acondor@chromium.org clank-performance-team@google.com Records the time spent executing RestrictedCookieManager::SetCookieFromString, plus its callback. Subsampled to 1/1000. See kHistogramSampleProbability in net/cookies/cookie_constants.h for details. Not reported for clients with low-resolution clocks. dylancutler@google.com src/net/cookies/OWNERS True iff the cookie set by RestrictedCookieManager is a partitioned cookie. Only logged when partitioned cookies are enabled and the cookie set by RCM will be accepted. Subsampled to 1/1000. See kHistogramSampleProbability in net/cookies/cookie_constants.h for details. morlovich@chromium.org src/net/cookies/OWNERS Logged on any document.cookie or CookieStore API operation. True if the browser-side site_for_cookies and the render-provided value match. Subsampled to 1/1000. See kHistogramSampleProbability in net/cookies/cookie_constants.h for details. Warning: Histogram was expired from 2025-03-09 to 2026-05-14. Data may be missing. morlovich@chromium.org src/net/cookies/OWNERS Logged on any document.cookie or CookieStore API operation. True if the browser-side top_frame_origin and the render-provided value match. Subsampled to 1/1000. See kHistogramSampleProbability in net/cookies/cookie_constants.h for details. Warning: Histogram was expired from 2025-01-26 to 2026-05-14. Data may be missing. tov@chromium.org src/net/OWNERS Captures to which port a connection to localhost was blocked. This only applies to the ports controlled by the RestrictAbusePortsOnLocalhost feature. Logged for every request that is initiated. tov@chromium.org src/net/OWNERS Captures to which restricted port a connection is initiated, regardless of whether this needs to be enforced. This only applies to the ports controlled by the RestrictAbusePorts feature. Logged for every request that is initiated. tju@google.com chrome-kids-eng@google.com Records if a URL classification in safe_search_api::URLChecker was provided from the cache or not. In the latter case, it distinguishes between true cache misses and cache hits which were unfit for classification. Missing data between 2024-04-28 and 2024-07-30. suzukikeita@chromium.org blink-network-stack@google.com The network error code of the session creation for preconnect to GoogleSearch using Quic for {JobType} job type. This is recorded everytime a preconnect is attempted to GoogleSearch, after preconnect attempt is finished. suzukikeita@chromium.org blink-network-stack@google.com Records whether the session is already established for a preconnect request to GoogleSearch using {ProtocolType}. This is recorded everytime a preconnect is attempted to GoogleSearch, after preconnect attempt is finished. suzukikeita@chromium.org blink-network-stack@google.com Whether the session create to {HostType} using {ProtocolType} had any traffic annotation tag associated. {ProtocolType}. This is recorded everytime a session create is attempted to GoogleSearch, after the attempt is finished. suzukikeita@chromium.org blink-network-stack@google.com The traffic annotation tag for the session create to {HostType} using {ProtocolType}. This is recorded everytime a session create is attempted to GoogleSearch, after the attempt is finished. wbjacksonjr@chromium.org src/net/extras/shared_dictionary/OWNERS Tracks the outcome of a network transaction only when a matching shared dictionary was found and advertised to the server in the request headers. It is emitted after the server's response headers have been received. The metric records whether the server actually used the dictionary for compression, which type of compression was used, or whether it chose not to use the dictionary. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS The maximum total size of dictionaries that can be stored in the Shared Dictionary cache, calculated based on available disk space in the browser process and notified to the network service. This is recorded once per a netwrok context only if the DiskCacheSize enterprise policy is not set. The original histogram "CacheMaxSize" didn't correctly recorded the actual maximum total size. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS The size of shared dictionary written on disk. Emitted when SharedDictionaryManagerOnDisk succeeded to store the dictionary in disk cache and database. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS The number of dictionaries evicted as there is no matching disk cache entry. Emitted when MismatchingEntryDeletionTask finishes. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS The number of disk cache entries evicted as the key is invalid for token. Emitted when MismatchingEntryDeletionTask finishes. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS The number of disk cache entries evicted as there is no matching metadata. Emitted when MismatchingEntryDeletionTask finishes. Warning: this histogram was expired from 2024-06-29 to 2024-08-27; data may be missing. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS The maximum total size of dictionaries that can be stored in the Shared Dictionary cache, as specified by an enterprise policy. This is recorded only once at a network context startup, and only if the DiskCacheSize enterprise policy is set to non zero value. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS The total count of shared dictionary on disk. Emitted when SharedDictionaryManagerOnDisk succeeded to store a dictionary in disk cache and database. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS The total size of shared dictionary on disk. Emitted when SharedDictionaryManagerOnDisk succeeded to store a dictionary in disk cache and database. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS The time taken to open a disk cache entry of a shared dictionary. Recorded when the entry is opened in the SharedDictionaryOnDisk. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS The time taken to read the data of a shared dictionary's disk cache entry. Recorded when the whole data is read in the SharedDictionaryOnDisk. wbjacksonjr@chromium.org src/net/extras/shared_dictionary/OWNERS After a dictionary is fetched, this metric tracks the outcome of the storage process. It records whether the dictionary was successfully written to the disk cache and its metadata was registered in the database, along with a variety of specific failure modes (e.g., disk write errors, database errors, size limit exceeded). pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS Records whether the shared dictionary metadata has been read when GetDictionary() is called for the first time. wbjacksonjr@chromium.org src/net/extras/shared_dictionary/OWNERS Records whether the shared dictionary metadata has been read when GetDictionary() is called for the first time, specifically for storages that were previously evicted. wbjacksonjr@chromium.org src/net/extras/shared_dictionary/OWNERS Records whether the shared dictionary metadata has been read when GetDictionary() is called for the first time, specifically for storages that were previously evicted due to memory pressure. wbjacksonjr@chromium.org src/net/extras/shared_dictionary/OWNERS Records the specific reason a Shared Dictionary Storage instance was previously evicted from the in-memory cache. This metric is emitted when `GetDictionarySync()` is called for the first time on a re-created storage instance for the same site isolation key. It provides insight into the causes of dictionary unavailability when a storage needs to be reloaded, helping to understand the real-world effectiveness of the feature, especially on memory-constrained devices. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS The time taken to read shared dictionary metadata from the database. Recorded when SharedDictionaryStorageOnDisk receives the result. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS The count of shared dictionary per top frame site on metadata. Emitted when SQLitePersistentSharedDictionaryStore is writing a dictionary in the database. Warning: this histogram was expired from 2024-12-20 to 2024-08-27; data may be missing. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS The size of shared dictionary per top frame site on metadata. Emitted when SQLitePersistentSharedDictionaryStore is writing a dictionary in the database. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS Records the status (OK or error code) returned by SQLitePersistentSharedDictionaryStore. Recorded when the database operation finishes. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS The delay to read a shared dictionary from a SharedDictionaryTransaction which is already aborted. Recorded when the dictionary is read, but the SharedDictionaryTransaction was aborted before the read operation was completed. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS The delay of SharedDictionaryTransaction caused by reading a shared dictionary for a {RequestType}. Recorded when the dictionary is read in the SharedDictionaryTransaction. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS Positive net error codes returned from the network layer for requests with Sec-Available-Dictionary header. Recorded when a SharedDictionaryNetworkTransaction received the result from the network layer. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS Records whether the response used a shared dictionary when it was available. This is recorded only for main frame requests. Recorded when the URLLoader request is completed. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS Positive net error codes of responses which used shared dictionaries. Recorded when the URLLoader request is completed. pmeenan@chromium.org src/net/extras/shared_dictionary/OWNERS The size of shared dictionary written in memory. Emitted when SharedDictionaryWriterInMemory succeeded to store the dictionary in memory. Warning: this histogram was expired from 2024-12-20 to 2024-08-27; data may be missing. fdoray@chromium.org ricea@chromium.org The outcome of enabling the FILE_SKIP_COMPLETION_PORT_ON_SUCCESS option on a TcpSocketIoCompletionPortWin. Recorded for 1/1000 attempts to set the option. pasko@chromium.org src/net/OWNERS The time it takes to execute SocketPosix::DoRead. Emitted at the end of DoRead. Subsampled to 1/1000. A few words on Rationale. This histogram can be used as an exemplary task duration on the network thread. Sometimes it will take several milliseconds to do a socket read due to OS state, like thrashing, or just randomly descheduling the task from the CPU. This DoReadDuration will show how often these events happen across the fleet. Regardless of its slowness in the field, the assumption is that the read duration cannot be realistically improved in the app. In other words, this duration is 'good enough' for any task on this thread. This will be a useful baseline when compared to duration of other tasks on the network thread, such as OnHandshakeIOCompleteDuration. dschinazi@chromium.org src/net/OWNERS The size of a socket's receive buffer when the attempt to change it via setsockopt failed. dschinazi@chromium.org src/net/OWNERS The size of a socket's send buffer when the attempt to change it via setsockopt failed. bashi@chromium.org src/net/OWNERS The counts of the flow control state of each frame (with stream and session flow control on). bashi@chromium.org src/net/OWNERS The percent compression achieved when compressing HEADERS frames. Warning: this histogram was expired from 2024-03-17 to 2024-04-24; data may be missing. bashi@chromium.org src/net/OWNERS Status of checking if a SPDY domain can handle a IP match. If a match is found, we successfully used the IP Pooling. If a match is not found, we could have used IP Pooling, except the TLS Cert didn't match the IP-pooled domain. dschinazi@chromium.org src/net/OWNERS HTTP status codes received in HTTP/2 HEADERS frames. If a stream contains one or more informational (1xx) response preceeding the main response, each status code is logged separately. bashi@chromium.org src/net/OWNERS Records whether the HTTP/2 session received an Accept-CH entry for the given origin, which GetAcceptChViaAlps() returns. Logged every time SpdyHttpStream::GetAcceptChViaAlps() is called, typically once for each request that is served over HTTP/2. bnc@chromium.org src/net/OWNERS Records both whether there are any valid origin fields and whether there are any invalid ones in ACCEPT_CH frames. Logged once for each TLS handshake of a HTTP/2 connection if ALPS is negotiated, regardless of the number of ACCEPT_CH frames received in ALPS. bnc@chromium.org src/net/OWNERS Records the status (OK or error code) returned by AlpsDecoder when parsing ALPS data. Logged once for each TLS handshake of a HTTP/2 connection if ALPS is negotiated. arichiv@chromium.org src/net/OWNERS Net.SpdySession.AlpsDecoderStatus records errors returned by the AlpsDecoder class in HTTP/2 responses. Errors logged to that histogram cause a session termination, which isn't always desired. This histogram records errors that we want to note, but that shouldn't stop the connection. bnc@chromium.org src/net/OWNERS Records the total number of setting parameters received in SETTINGS frames in ALPS. Logged once for each TLS handshake of a HTTP/2 connection if ALPS is negotiated, regardless of the number of SETTINGS frames received in ALPS. bashi@chromium.org src/net/OWNERS Net error codes when SpdySession was closed. bashi@chromium.org src/net/OWNERS Socket connected status in SpdySession::CreateStream. hayato@chromium.org src/net/OWNERS Records the net error code returned by GetPeerAddress inside SpdySession::GetRemoteEndpoint. This is recorded every time GetRemoteEndpoint is called. bashi@chromium.org src/net/OWNERS Records the error code value received in each HTTP/2 GOAWAY frame. bashi@chromium.org chrome-network-stack@google.com Records the number of {StreamState} streams when a SpdySession receives an HTTP/2 SETTINGS frame for the first time. bashi@chromium.org src/net/OWNERS Records the error code value received in each HTTP/2 RST_STREAM frame. bashi@chromium.org src/net/OWNERS Whether the server has advertised support for WebSockets over HTTP/2. bashi@chromium.org src/net/OWNERS The type of SPDY Protocol error encountered. bashi@chromium.org src/net/OWNERS WARNING: r181910 added an enum value in the middle, so don't trust the counts for values 9 and above for Chrome builds after that revision. The type of SPDY Protocol error encountered when talking to a google.com server. dschinazi@chromium.org src/net/OWNERS The type of SPDY Session used when looking up a session. dschinazi@chromium.org src/net/OWNERS Whether SpdySession::Get{Peer,Local}Address was called when the connection had no socket. bashi@chromium.org src/net/OWNERS The number of pushed, but abandoned streams over a single session. bashi@chromium.org src/net/OWNERS The number of streams issued over a single session. horo@chromium.org src/net/OWNERS Records the size of the database file of the SQL-based disk cache. This is recorded when the backend is initialized. horo@chromium.org src/net/OWNERS Records the number of entries in the SQL-based disk cache. This is recorded when the backend is initialized. horo@chromium.org src/net/OWNERS Records the result of the incremental vacuum attempt of the SQL-based disk cache backend. This helps track the frequency of different failures and aborts. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Measures the number of pages vacuumed during an incremental vacuum attempt of the SQL-based disk cache backend. This metric is recorded immediately after the incremental vacuum attempt {ResultType}. It helps diagnose performance regressions. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Measures the time taken to run the incremental vacuum of the SQL-based disk cache backend. This metric is recorded immediately after the incremental vacuum attempt {ResultType}. It helps diagnose performance regressions. The histogram is not reported on Windows machines without high resolution clocks. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Records the location where an in-memory index mismatch is detected in the SQL-based disk cache. This helps identify the operations that are most prone to causing inconsistencies between the in-memory index and the on-disk database. horo@chromium.org src/net/OWNERS Measures the time taken to load the in memory index of the SQL-based disk cache backend. This metric is recorded immediately after the in-memory index is loaded from the database. The histogram is not reported on Windows machines without high resolution clocks. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Records the configured maximum size of the SQL-based disk cache. This is recorded when the backend is initialized. horo@chromium.org src/net/OWNERS Errors reported by SQLite while initializing or using the SQL-based disk cache backend database. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Records the total size of entries in the SQL-based disk cache. This is recorded when the backend is initialized. horo@chromium.org src/net/OWNERS Measures the number of pages in the WAL file when the SQL-based disk cache checkpoint is performed. This metric is recorded immediately after the checkpoint succeeded or failed. It helps diagnose performance regressions. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Measures the time taken to run the checkpoint of the SQL-based disk cache backend. This metric is recorded immediately after the checkpoint succeeded or failed. It helps diagnose performance regressions. The histogram is not reported on Windows machines without high resolution clocks. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Records the number of entries deleted during a single eviction run in the SQL-based disk cache. This is recorded after the eviction is finished. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Records the number of entries scanned in the database or in-memory index during a single eviction candidate selection run in the SQL-based disk cache. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Records the SQLite error code when SelectEvictionCandidates fails during a single eviction run in the SQL-based disk cache. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Records the time taken to select entries for deletion during a single eviction run in the SQL-based disk cache. This is recorded after the selection of the candidates is finished. The histogram is not reported on Windows machines without high resolution clocks. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Measures the time delay from posting a task to its execution for the method of the SQL-based disk cache backend. This metric is recorded immediately before the logic for the method is executed. It helps diagnose potential delays caused by task queueing. The histogram is not reported on Windows machines without high resolution clocks. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Records the result of the method of the SQL-based disk cache backend. This helps track the frequency of different failures. This histogram is not reported if any data corruption was detected during the operation. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Records the error result of a method of the SQL-based disk cache backend when a data corruption is detected. This helps track the frequency of different failures in corrupted scenarios. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Measures the time taken to run the method of the SQL-based disk cache backend. This metric is recorded immediately after the logic for the method succeeded or failed. It helps diagnose performance regressions. This histogram is not reported if any data corruption was detected during the operation. The histogram is not reported on Windows machines without high resolution clocks. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Records the number of doomed entries deleted during a single call to DeleteDoomedEntries. This operation is used for garbage collection of orphaned cache entries, which can be left behind after events like a network service crash. It deletes all entries marked as doomed, except for a specified set of entries that are still actively in use. Recorded each time the cleanup operation is run. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Records the percentage of live cache entries that have associated hints set. This is recorded once each time the in-memory index is loaded from the database. This data is used to evaluate the efficiency of the in-memory data structures. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Measures the time an exclusive operation spends in the queue before it is run in the SQL-based disk cache. This helps monitor potential delays and starvation issues for exclusive operations. Recorded when an exclusive operation is run. The histogram is not reported on Windows machines without high resolution clocks. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Records the error status that can occur during the fake index file check of the SQL-based disk cache. This check is performed during the cache initialization. horo@chromium.org src/net/OWNERS Measures the total duration of the SQL disk cache initialization. Recorded when the initialization completes {SuccessOrFailure}. The histogram is not reported on Windows machines without high resolution clocks. horo@chromium.org src/net/OWNERS Records whether all memory entry data hints loaded from the SQL-based disk cache database are valid (fit within the expected 2-bit range). This is recorded once when the in-memory index is loaded from the database. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Measures the time a normal operation spends in the queue before it is run in the SQL-based disk cache. This helps monitor potential delays and starvation issues, especially when exclusive operations are frequent. Recorded when a normal operation is run. The histogram is not reported on Windows machines without high resolution clocks. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Records the result of an optimistic write operation to the SQL disk cache. This is recorded when the optimistic write completes on the background thread, indicating whether the underlying persistent store operation succeeded or failed. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org src/net/OWNERS Records whether a disk cache write operation was performed optimistically (synchronously) or not. True indicates the write was optimistic, false indicates it was queued for asynchronous processing. This metric will be emitted every time a cache entry write happens. Not recorded when kSqlDiskCacheReduceUma is true. horo@chromium.org net-dev@chromium.org The result of {Method} in SqlSharedCacheIndexDatabase. horo@chromium.org net-dev@chromium.org The processing time of {Method} in SqlSharedCacheIndexDatabase. Recorded on {TimeType}. The histogram is not reported on Windows machines without high resolution clocks. davidben@chromium.org trusty-transport@chromium.org For each TLS connection that offered Encrypted ClientHello (ECH), what was the result of that connection. In particular, this measures if the recovery flow was needed and, if so, whether it fixed the connection. davidben@chromium.org estark@chromium.org trusty-transport@chromium.org For each TLS connection, recorded when the connection attempt completes, whether the trust anchor IDs retry flow was triggered and, if so, whether it fixed the connection. agl@chromium.org rsleevi@chromium.org The SSL/TLS cipher suite that was negotiated. Recorded for each SSL/TLS connection in the socket pool where Connect() succeeds. agl@chromium.org Counts of specific error codes returned when opening an SSL connection. davidben@chromium.org estark@chromium.org trusty-transport@chromium.org Counts of specific error codes returned when opening an SSL connection. Recorded for connections to servers that advertise support for Trust Anchor IDs in DNS. agl@chromium.org Time from when the Connect() starts until it completes. davidben@chromium.org hchao@chromium.org chrome-secure-web-and-net@chromium.org Time from when the Connect() starts until it completes. Recorded for successful connections to servers that used the Server Padding extension. davidben@chromium.org estark@chromium.org trusty-transport@chromium.org Time from when the Connect() starts until it completes. Recorded for successful connections to servers that advertise support for Trust Anchor IDs in DNS. agl@chromium.org rsleevi@chromium.org sigbjorn@opera.com The curve used for key exchange for ECDHE. Recorded for each SSL/TLS connection in the socket pool where Connect() succeeds. davidben@chromium.org src/net/OWNERS For each non-resumption TLS handshake that sends a client certificate, the signature algorithm used by the client. Warning: this histogram was expired from 2025-02-10 to 2025-06-25; data may be missing. pasko@chromium.org src/net/ssl/OWNERS The time it takes to execute SSLClientSocketImpl::OnHandshakeIOComplete. Measured at the end of OnHandshakeIOComplete. Subsampled to 1/1000. davidben@chromium.org src/net/OWNERS For each TLS handshake, whether the server supports the Extended Main Secret extension (RFC 7627). If the server negotiated TLS 1.3, which is always patched against this bug, it is considered to support the extension. Servers that do not implement this extension may be vulnerable to attacks when using channel binding or key exporters. Note this extension was originally defined with a different name in RFC 7276. It is being renamed in draft-ietf-tls-rfc8446bis to Extended Main Secret. With both names, it is often referred to by the acronym EMS. davidben@chromium.org src/net/OWNERS For each successful TLS handshake, what kind of handshake was used. This metric distinguishes TLS 1.2 (or earlier) and TLS 1.3 as they have very different resumption and round-trip behaviors. It also distinguishes full handshakes, resumption, 0-RTT, and False Start. bashi@chromium.org trusty-transport@chromium.org Indicates whether a TLS 1.3 connection ended up using 0-RTT or not, and why; this includes reasons such as the server declining to resume the connection and the client not having enough tickets available. This metric is only recorded if 0-RTT supported is enabled in the client. Warning: this histogram was expired from 2022-05-11 to 2024-11-05; data may be missing. bashi@chromium.org trusty-transport@chromium.org Indicates whether a TLS 1.3 connection to a Google host ended up using 0-RTT or not, and why; this includes reasons such as the server declining to resume the connection and the client not having enough tickets available. This metric is only recorded if 0-RTT supported is enabled in the client. Warning: this histogram was expired from 2022-05-11 to 2024-11-05; data may be missing. cthomp@chromium.org trusty-transport@chromium.org Counts when the SSLKEYLOGFILE environment variable or --ssl-key-log-file command-line flag are set, and when they enable the SSLKeyLogger feature. This histogram was temporarily expired from M92 until M104. dschinazi@chromium.org src/net/OWNERS For each TLS handshake, whether ALPN was negotiated; and if so, the negotiated protocol. davidben@chromium.org src/net/OWNERS For each TLS handshake, whether the server supports the renegotiation_info extension (RFC 5746). If the server negotiated TLS 1.3, which removed renegotiation, it is considered to support the extension. Servers that do not support it are potentially vulnerable to an attack involving TLS renegotiation. davidben@chromium.org src/net/OWNERS For each SSL connection with a full handshake using a DHE- or ECDHE-based key exchange, the signature algorithm used to authenticate the peer. In TLS 1.2, this is the signature on the ServerKeyExchange message. (Note: Although the ECDSA values specify a curve, the curve is only enforced in TLS 1.3.) davidben@chromium.org The SSL/TLS version that was negotiated. Recorded for each SSL/TLS connection in the socket pool where Connect() succeeds. apaseltiner@chromium.org net-dev@chromium.org True when strict base64 decoding of a Structured Header succeeds, false when it fails but non-compliant lenient decoding succeeds. Emitted whenever a byte-sequence-typed bare item (https://www.rfc-editor.org/rfc/rfc8941.html#name-byte-sequences) is successfully parsed, regardless of whether subsequent parsing of the remainder of the top-level Structured Header is successful. See crbug.com/393153699 and crbug.com/393408763 for details. apaseltiner@chromium.org net-dev@chromium.org Whether a decimal point in a Structured Header is erroneously accepted without any digits following it. Emitted whenever a decimal-typed bare item (https://www.rfc-editor.org/rfc/rfc8941.html#name-decimals) is successfully parsed, regardless of whether subsequent parsing of the remainder of the top-level Structured Header is successful. See crbug.com/517189418 for details. apaseltiner@chromium.org net-dev@chromium.org The result (success or failure) of parsing a structured-header {Type}. See crbug.com/377941140 for more details. apaseltiner@chromium.org net-dev@chromium.org The time taken to parse a structured-header {Type} via the net::structured_headers::Parse{Type} function. Recorded for all clients, not just ones with high-resolution time, and recorded regardless of whether parsing succeeds or fails. See crbug.com/377941140 for more details. hayato@chromium.org blink-network-stack@google.com The request priority passed to net::GetTaskRunner. Recorded with 0.001 probability when net::GetTaskRunner is called. mmenke@chromium.org Time from when the Connect() starts until it completes. Only times under 10 minutes are logged. {DnsImpact2} bashi@chromium.org src/net/OWNERS Time taken when failed to establish a TCP connection to an individual endpoint. This includes the case when the attempt was aborted by a higher layer (likely because it took too long). bashi@chromium.org src/net/OWNERS Time taken to successfully establish a TCP connection to an individual endpoint. arichiv@chromium.org awillia@chromium.org katabolism-finch@google.com The time between the release of a socket on a particular local address/port and its future re-use in another connection. The time for the last release is cleared if the open is successful, and reused without overwriting if not. Recorded once for each attempted re-use of an address/port for TCP on Windows via TCPSocketWin::DoConnectComplete. This second version of the histogram uses UmaHistogramLongTimes100 where the original used UmaHistogramTimes, which was insufficient. {IPType} {Result} arichiv@chromium.org katabolism-finch@google.com Records when SocketPoolAdditionalCapacity is constructed with invalid constants, or when transitions are calculated with invalid parameters. Most errors cannot cause exits as the code path is critical, so we just log and take a low-risk route. This data is used to validate proper behavior of the probability function. See crbug.com/415691664 for more details. arichiv@chromium.org katabolism-finch@google.com Records the probability of a state transition returned when a socket is having action, {Action}, performed. This data is used to validate proper behavior of the probability function. See crbug.com/415691664 for more details. arichiv@chromium.org katabolism-finch@google.com Records the number of active sockets at the time the new state for the socket pool is calculated. The state {PastState} was the behavior before the action, {Action}, and the state {NextState} defines the behavior after. This data is used to validate proper behavior of the probability function. See crbug.com/415691664 for more details. svaldez@chromium.org timhuber@google.com yhirano@chromium.org The net error for a failed Fetch API call with an associated Trust Tokens operation. This was originally added to help debug a surfeit of 'TypeError: failed to fetch' observed in live testing, and it is now retained since it's useful for ongoing health monitoring. {TrustTokenOperationType} svaldez@chromium.org timhuber@google.com The //net status codes for resource requests that had started Trust Tokens operations. This captures the reasons why requests with Trust Tokens operations fail and might help debug a surfeit of "TypeError: failed to fetch" observed in live testing. {OperationResult} {TrustTokenOperationType} svaldez@chromium.org timhuber@google.com For Private State Token (PST) operations, the result of the operation. This is recorded everytime a PST operation completed in URLLoader. See network::URLLoader::OnDoneBeginningTrustTokenOperation for details. This can help debug the use of the API observed in the ecosystem. {TrustTokenOperationType} svaldez@chromium.org timhuber@google.com Time taken to execute various parts of a Trust Tokens operation: 1. Begin = time taken to execute the operation's outbound Begin half 2. Server = time from the end of the outbound Begin half to the beginning of the inbound Finalize half 3. Finalize = time taken to execute the inbound Finalize half 4. Total = Begin + Server + Finalize (only recorded for operations that see server responses) The other variants (see TrustTokenOperationType variants): 1. Outcome (Success or Failure) 2. The type of Trust Tokens operation 3. Whether the operation was platform-provided (i.e. executed by a method other than a direct HTTP request to the issuer's server) aykutb@chromium.org trust-token@google.com The crypto version used for Private State Token (PST) operations. This is recorded every time a PST issuance operation started. This will be used to remove the unused crypto versions. svaldez@chromium.org timhuber@google.com Whether the Trust Tokens redemption logic received an empty redemption request string from its underlying cryptographic library. If this is always true, but servers occasionally see empty redemption request headers, it might suggest middleware (namely extensions) is removing the headers' values. svaldez@chromium.org timhuber@google.com For resource requests that bear Trust Tokens parameters, the results of running TrustTokenRequestHelperFactory to initialize Trust Tokens handling logic. This might help debug a surfeit of failed Trust Tokens issuance operations observed in live testing. {TrustTokenOperationType} aymanm@google.com ricea@chromium.org src/net/OWNERS Records the number of UDP datagrams read and sliced from a UDP_GRO superpacket (or un-coalesced datagram) per read operation when UDP Generic Receive Offload (UDP_GRO) is enabled on a UDPSocketPosix. This metric is sampled at 1% of successful GRO read operations. This metric is only emitted if `ReadMultiple()` is called. aymanm@google.com ricea@chromium.org src/net/OWNERS Records the number of UDP datagrams read in a single batch read (recvmmsg) operation on a UDPSocketPosix. This metric is sampled at 1% of successful recvmmsg read operations. This metric is only emitted if `ReadMultiple()` is called and recvmmsg is used. aymanm@google.com ricea@chromium.org src/net/OWNERS Records the result of enabling UDP Generic Receive Offload (UDP_GRO) via setsockopt on a UDPSocketPosix. This metric is sampled at 1% of socket creations. dschinazi@chromium.org src/net/OWNERS The time spent in closesocket call in UDPSocketWin::Close. yyanagisawa@chromium.org chrome-loading@google.com We aim to offload the critical client hint check in the network layer. This metric records true if we still need to call the AcceptCHFrameObserver for checking the hint. It is recorded only if the received header has the ACCEPT_CH frame on the QUIC connection set up timing. Refer AcceptCHFrameInterceptor::OnConnected() for details. yyanagisawa@chromium.org chrome-loading@google.com Records the IPC round trip time to handle the AcceptCH frame in the browser process. It is recorded if there are hints in the ACCEPT_CH frame that were not included in the original request during URLLoader to handle a connection. It is time just before the OnAcceptCHFrameReceived() IPC call until the IPC callback is called. Refer AcceptCHFrameInterceptor::OnConnected() for details. Note: The metrics is reported from all clients. Some of clients may not have high resolution timer that can send metrics with microseconds precision. yyanagisawa@chromium.org chrome-loading@google.com Records if the received the ACCEPT_CH frame should be checked by the observer. It is recorded only if the received header has the ACCEPT_CH frame. Refer AcceptCHFrameInterceptor::OnConnected() for details. yyanagisawa@chromium.org chrome-loading@google.com Records the response status of the the AcceptCH frame handling in the browser process. It is recorded if there are hints in the ACCEPT_CH frame that were not included in the original request during URLLoader to handle a connection. It is time just before the OnAcceptCHFrameReceived() IPC call until the IPC callback is called. Refer AcceptCHFrameInterceptor::OnConnected() for details. yyanagisawa@chromium.org chrome-loading@google.com Records whether an ACCEPT_CH frame was received on an HTTP/2 connection. This is recorded when the URLLoader is connected. yyanagisawa@chromium.org chrome-loading@google.com Records whether an ACCEPT_CH frame was received on an HTTP/3 connection. This is recorded when the URLLoader is connected. bashi@chromium.org blink-network-stack@google.com Records whether a URLLoader has a TrustedHeaderClient when starting a transaction. Logged before the URLLoader starts a transaction. bashi@chromium.org blink-network-stack@google.com Records the duration of the TrustedHeaderClient::OnBeforeSendHeaders IPC call. Logged when the callback for OnBeforeSendHeaders is received in URLLoader::OnBeforeSendHeadersComplete. yoichio@chromium.org chrome-loading@google.com Records the proportion of writes when URLLoader was blocked by MOJO_RESULT_SHOULD_WAIT where the renderer process seems to be busy. Logged every time a URLLoader completes a load. yoichio@chromium.org chrome-loading@google.com Records the proportion of writes when URLLoader was blocked by MOJO_RESULT_SHOULD_WAIT where the renderer process seems to be busy. Logged every time a URLLoader completes a load and only when SlopBucket is actually used and response cache type is {CacheEntryStatus}. nidhijaju@chromium.org chrome-loading@google.com Time spent on the callback called when the response head is received. Logged every time a throttle receives a response. cduvall@chromium.org jam@chromium.org Measures time {Throttle} deferred a request on WillProcessResponse. Logged every time {Throttle} defers a request. cduvall@chromium.org jam@chromium.org Measures time {Throttle} deferred a request on WillStartRequest. Logged every time {Throttle} defers a request. cduvall@chromium.org jam@chromium.org Measures time a URLLoaderThrottle was deferred when on {Stage}. Logged every time a throttle defers a request. cduvall@chromium.org jam@chromium.org Measures time a URLLoaderThrottle took to execute {Stage}. Logged every time a throttle runs. davidvc@chromium.org kaustubhag@chromium.org Records for each URLRequest whether its referrer contains more information than just its origin (this is the case exactly when it has at least one character apart from a leading slash in its path, e.g. http://example.com/a but not http://example.com/ nor http://example.com). {ReferrerPolicySameOrigin} davidvc@chromium.org kaustubhag@chromium.org Records the distribution of referrer policies provided with outgoing URLRequests, and whether or not the requests were cross-origin. {ReferrerPolicySameOrigin} ricea@chromium.org blink-network-stack@google.com Positive net error codes that WebSockets end with, including OK and ABORTED. bashi@chromium.org blink-network-stack@google.com Fallback result of WebSocket handshakes. Recorded when a WebSocket handshake completed. yhirano@chromium.org ricea@chromium.org src/net/OWNERS Results of WebSocket handshakes, including information whether the handshake happened over HTTP/1, HTTP/2 or HTTP/3. Use this histogram as a baseline for investigating feature usage counters. ricea@chromium.org patilmayur@microsoft.com blink-network-stack@google.com Whether a WebSocket-over-HTTP/3 handshake can reuse an existing QUIC session that supports Extended CONNECT. Recorded for WebSocket handshakes when QUIC and WebSocket-over-HTTP/3 are enabled and the proxy is direct. ricea@chromium.org blink-network-stack@google.com The size of the value of the Sec-WebSocket-Protocol header in the WebSocket handshake request, in bytes. Recorded once for every WebSocket handshake request sent, regardless of HTTP version. yhirano@chromium.org ricea@chromium.org All HTTP status codes seen during WebSocket handshakes. bashi@chromium.org blink-network-stack@google.com Negotiated HTTP Datagram version. Recorded when a WebTransport session becomes ready. bashi@chromium.org blink-network-stack@google.com Negotiated WebTransport version. Recorded when a WebTransport session becomes ready. nidhijaju@chromium.org blink-network-stack@google.com Compressed/Decompressed size ratio as a percentage when using zstd. nidhijaju@chromium.org blink-network-stack@google.com Error code observed after failing to decode a Zstd-compressed response stream from the network. Recorded once per layer of Zstd compression that encountered a decoding error. See Net.ZstdFilter.Status to get the overall error rate. Warning: this histogram was expired from 2024-12-31 to 2025-06-25; data may be missing. nidhijaju@chromium.org blink-network-stack@google.com The maximum memory allocation by the zstd library to decode a Zstd-compressed response stream from the network. Recorded once per layer of Zstd compression (usually once per zstd-compressed response), at the end of decoding. nidhijaju@chromium.org blink-network-stack@google.com The final state of the ZstdFilter. Recorded once per layer of Zstd compression (usually once per zstd-compressed response), at the end of decoding. Warning: this histogram was expired from 2025-05-05 to 2025-06-25; data may be missing. suzukikeita@chromium.org blink-network-stack@google.com The session creation initiator for a {ProtocolType} to GoogleSearch. This is recorded for each session with at least one stream created, when closing the session to GoogleSearch, and when {Used}. morlovich@chromium.org wanderview@chromium.org This even is emitted when SimpleCache is forced to either close some files due to being over FD limit, or reopen files it had to close due to such limit earlier. Warning: this histogram was expired from 2022-11-13 to 2025-10-17; data may be missing. jkarlin@chromium.org shivanisha@chromium.org The size of the cache at the time that the index has finished initializing. This is the disk space actually used. {CacheInstance} wanderview@chromium.org morlovich@chromium.org yoichio@chromium.org chrome-loading@google.com Record the outcome of the simple DiskCache consistency check. The consistency check is executed every time a simple DiskCache is initialized. This check includes both inspecting the schema on disk and upgrading the schema to the latest version if necessary. {CacheInstance} wanderview@chromium.org morlovich@chromium.org Record the final consistency check value after index files have been deleted from an empty simple DiskCache. This histogram will tell us how often the index deletion code results in a functioning cache. It will also tell us what kind of consistency failures cannot be recovered due to index file deletion. {CacheInstance} morlovich@chromium.org wanderview@chromium.org The time from the creation of the simple cache backend until the index has been loaded from disk. {CacheInstance} morlovich@chromium.org wanderview@chromium.org The time from the creation of the simple cache backend until the index fails to load. {CacheInstance} wanderview@chromium.org morlovich@chromium.org Record if any simple DiskCache index files were deleted from an otherwise empty cache after the cache failed the initialization consistency check. This histogram will tell us how often the deletion code is triggering and modifying the disk. {CacheInstance} Warning: this histogram was expired from 2023-12-08 to 2025-10-17; data may be missing. morlovich@chromium.org wanderview@chromium.org The time to close (write out footers and close the file) the cache entry on disk. Includes only successful operations. {CacheInstance} morlovich@chromium.org wanderview@chromium.org The time to access (open for creation and write initial headers) the cache entry on disk. Includes only successful operations. {CacheInstance} morlovich@chromium.org wanderview@chromium.org The time to remove the cache entry on disk. Includes only individually doomed entries, not entries bulk-erased during eviction. {CacheInstance} morlovich@chromium.org wanderview@chromium.org The time to access (open and read headers from all files) the cache entry on disk. Includes only successful operations. {CacheInstance} morlovich@chromium.org wanderview@chromium.org The time to write a chunk to the cache entry on disk. Includes only successful operations. {CacheInstance} morlovich@chromium.org wanderview@chromium.org The time, in ms, spent creating a new entry on disk. This measures time from when the method got removed from SimpleEntryImpl's operation queue till when the callback is about to be invoked, and includes all Open and Create operations, despite what the name may suggest. {CacheInstance} morlovich@chromium.org wanderview@chromium.org The number of entries to be erased in an eviction. {CacheInstance} morlovich@chromium.org wanderview@chromium.org Time spent completing an eviction. {CacheInstance} morlovich@chromium.org wanderview@chromium.org Time spent selecting entries for eviction. {CacheInstance} morlovich@chromium.org wanderview@chromium.org The size of the header stream of a Simple Cache entry, emitted every time the headers are written or rewritten. {CacheInstance} morlovich@chromium.org wanderview@chromium.org The number of entries in a newly created index file. {CacheInstance} morlovich@chromium.org wanderview@chromium.org The state the index file is at when an attempt is made to load from it. {CacheInstance} wanderview@chromium.org morlovich@chromium.org The method used to initialize the simple cache index. {CacheInstance} jkarlin@chromium.org morlovich@chromium.org The number of entries in the index at the time that the index has finished initializing. {CacheInstance} morlovich@chromium.org wanderview@chromium.org Time (as measured on the worker pool) spent restoring the index file by iterating directory entries. {CacheInstance} yoichio@chromium.org chrome-loading@google.com The result of writing the index file to disk. Recorded each time the index is written to disk. {CacheInstance} morlovich@chromium.org jkarlin@chromium.org The size limit computed for the cache. {CacheInstance} morlovich@chromium.org wanderview@chromium.org At the time that an entry is opened, the state of that entry in the index. {CacheInstance} wanderview@chromium.org morlovich@chromium.org Record the original consistency error produced by a simple DiskCache before a retry was attempted and succeeded. This histogram will tell us what kind consistency errors can be recovered from in practice. {CacheInstance} wanderview@chromium.org morlovich@chromium.org Record the outcome of the simple DiskCache consistency check after retrying initialization on an empty cache directory. The retry may occur after corrupt index files were deleted from an empty cache. The retry may also occur if the original failed initialization left an empty directory without any index files. This histogram will tell us how often the retry succeeds in recovering caches and also which consistency checks cannot be recovered. {CacheInstance} morlovich@chromium.org wanderview@chromium.org Count of the number of entries recorded in the index, but not actually present in the cache. Recorded each time a stale index is found and a directory rescan is required. {CacheInstance} morlovich@chromium.org wanderview@chromium.org Count of the number of entries present in a cache, but not recorded in the index. Recorded each time a stale index is found and a directory rescan is required. {CacheInstance} morlovich@chromium.org wanderview@chromium.org The quality of a "stale" index compared to a full directory rescan. Recorded immediately after every directory scan for stale indexes. {CacheInstance} ashleynewson@chromium.org morlovich@chromium.org The kind of prefetching performed, if any, when executing SimpleSynchronousEntry::ReadAndValidateStream0AndMaybe1(). {CacheInstance} Warning: this histogram was expired from 2023-12-10 to 2025-10-17 and 2026-01-13 to 2026-07-07; data may be missing.