summaryrefslogtreecommitdiffstats
path: root/container-core/src/main/java/com/yahoo/container/jdisc
Commit message (Collapse)AuthorAgeFilesLines
...
* Merge pull request #16227 from vespa-engine/gjoranv/remove-StateMetricConsumergjoranv2021-01-295-173/+25
|\ | | | | Gjoranv/remove state metric consumer
| * Remove the StateMetricConsumer class and its factory.gjoranv2021-01-262-56/+0
| |
| * Remove usage of MetricsPresentationConfig.slidingwindowgjoranv2021-01-261-5/+1
| |
| * Update package-info description.gjoranv2021-01-261-7/+5
| |
| * Remove metrics handling from the StateMonitor class.gjoranv2021-01-263-100/+10
| |
| * Handle null dimensions without throwing NPE.gjoranv2021-01-252-7/+11
| |
| * Revert "Access log optimizations [run-systemtest]"Morten Tokle2021-01-221-23/+83
| |
* | Deprecate LoggingRequestHandler constructors taking AccessLogBjørn Christian Seime2021-01-211-22/+33
|/ | | | | Add replacement constructors without AccessLog. Remove use of deprecated constructors for internal handlers.
* Replace AccessLogEntry with non-blocking RequestLogEntryBjørn Christian Seime2021-01-211-83/+23
| | | | | | | Keep AccessLogEntry as interface for adding extra information in handlers, but use the new RequestLogEntry for access log serialization. Introduce new interface RequestLog that AccessLog class implements (to simplify unit testing). Rename AccessLogInterface to RequestLogHandler. Remove unused class AccessLogSampler.
* Non-functional-changes-onlyJon Bratseth2020-11-1817-108/+90
|
* Merge pull request #14805 from ↵Henning Baldersheim2020-10-131-55/+32
|\ | | | | | | | | vespa-engine/hakonhall/take-statemonitor-snapshot-every-60s Take StateMonitor snapshot every 60s
| * Avoid mock thread factory and simplify constructionHåkon Hallingstad2020-10-101-35/+26
| |
| * Take StateMonitor snapshot every 60sHåkon Hallingstad2020-10-101-33/+19
| | | | | | | | | | | | | | | | | | Using Executors simplifies the code and fixes the following 2 small problems: - A tiny drift of the 1m interval: It starts the next snapshot 1m + the time before wait() return + the time until currentTimeMillis(). - May potentially (but unlikely) invoke wait() with negative (throws exception) or 0 argument (waits forever): There is no test on the returned long from currentTimeMillis().
* | Set RequestType both waysJon Bratseth2020-10-092-3/+5
|/
* Use OptionalJon Bratseth2020-10-071-6/+6
|
* Let handlers dedfine a default request typeJon Bratseth2020-10-011-2/+18
|
* Merge pull request #14634 from vespa-engine/jonmv/async-feed-handlerJon Marius Venstad2020-10-012-29/+5
|\ | | | | Async document v1
| * Add equals and hashCode to some classesJon Marius Venstad2020-09-301-1/+1
| |
| * Revert "Revert "Jonmv/async doc v1 implementation""Jon Marius Venstad2020-09-301-28/+4
| | | | | | | | This reverts commit c6aded1606112a54969f56403085ca90d61dac8f.
| * Revert "Jonmv/async doc v1 implementation"Harald Musum2020-09-291-4/+28
| |
| * Merge pull request #14554 from vespa-engine/jonmv/async-doc-v1-implementationJon Marius Venstad2020-09-291-28/+4
| |\ | | | | | | Jonmv/async doc v1 implementation
| | * Common metrics code and abi specJon Marius Venstad2020-09-281-5/+2
| | |
| | * Eliminate code duplication in generating metric context for requestsJon Marius Venstad2020-09-251-23/+2
| | |
* | | Change access modifier to 'public'Bjørn Christian Seime2020-09-281-2/+2
|/ /
* / Provide underlying executor through getterBjørn Christian Seime2020-09-281-0/+2
|/
* Non-functional changesJon Marius Venstad2020-09-252-5/+3
|
* Allow error response for worker pool overload to be customized by handlerBjørn Christian Seime2020-09-211-5/+12
|
* Merge pull request #14454 from vespa-engine/jonmv/async-document-v1Jon Marius Venstad2020-09-182-5/+6
|\ | | | | Jonmv/async document v1
| * Non-functionalJon Marius Venstad2020-09-142-5/+6
| |
* | No functional changesJon Bratseth2020-09-181-2/+2
| |
* | Propagate requestType through jdisc core responseJon Bratseth2020-09-142-5/+2
|/
* Allow setting a request type explicitlyJon Bratseth2020-09-022-12/+18
| | | | | | This lets handler authors control the requestType explicitly by setting it on the HttpResponse, which is useful to avoid misclassification of POST requests to reading handlers as writes.
* Avoid code duplication and make more clear that close is simply flushJon Marius Venstad2020-08-311-5/+1
|
* Time out connections on the IOThread levelJon Bratseth2020-08-281-14/+4
| | | | | | | Time out connections on the IOThread level instead of leaving this to Apache. Keep old connections alive for a while after timeout and keep polling them such that, if the old connection hits a different real behind a VIP than the new connection we'll still get the replies.
* Improvements to handling of cluster removalJon Bratseth2020-07-021-6/+9
| | | | | | | | | - Don't change health status to "initializing" when creating a new VipStatus, as 'initializing' now requires all clusters to be up to transition to 'up', which means that if we're already up but are missing a cluster we'll go from 'up' to 'initializing' and stay there. - Forget up/down status for removed clusters. - Nicer logging on ignorable reconfiguration errors.
* Use correct log Level class where search & replace has failed.gjoranv2020-04-251-1/+1
|
* Replace remaining LogLevel.<level> with corresponding Levelgjoranv2020-04-251-1/+1
|
* Map remaining DEBUG/SPAM/ERROR/FATAL -> Level.FINE/FINEST/SEVEREgjoranv2020-04-251-2/+2
|
* LogLevel -> Level for isLoggable()gjoranv2020-04-252-3/+3
|
* LogLevel.ERROR -> Level.SEVEREgjoranv2020-04-251-1/+1
|
* LogLevel.WARNING -> Level.WARNINGgjoranv2020-04-252-3/+3
|
* LogLevel.INFO -> Level.INFOgjoranv2020-04-251-1/+1
|
* LogLevel.DEBUG -> Level.FINEgjoranv2020-04-252-3/+3
|
* Import java.util.logging.Level instead of com.yahoo.log.LogLevelgjoranv2020-04-255-5/+5
|
* Update method and constant names to match new metric nameBjørn Christian Seime2020-03-131-2/+2
|
* Use plural ending in new metric namesBjørn Christian Seime2020-03-131-1/+1
|
* Introduce metric for unhandled exceptions from request handlersBjørn Christian Seime2020-03-121-0/+3
|
* Share common metric dimensions with package-local subclassesBjørn Christian Seime2020-03-121-1/+4
|
* Add scheme, port and handler-name as dimensionsBjørn Christian Seime2020-02-201-0/+6
|
* If debugging of SearchHandler is turned on, the 1k first request will have ↵Henning Baldersheim2020-02-111-2/+2
| | | | | | trace and timing information. The 1k number is configurable in the container-http config.