| b845c4f | Brian Seeders | 22 March 2024, 17:38:17 UTC | Just always use 4 workers | 22 March 2024, 17:38:17 UTC |
| 3ce919d | Brian Seeders | 22 March 2024, 16:50:40 UTC | Try much lower worker count | 22 March 2024, 16:50:40 UTC |
| 126a5d5 | Brian Seeders | 22 March 2024, 16:48:39 UTC | Merge branch 'main' into bk-cost-opt-sizing | 22 March 2024, 16:48:39 UTC |
| ecb4223 | Mark Tozzi | 22 March 2024, 16:48:00 UTC | [ESQL] Migrate PropagateEquals optimization (#106627) Relates to #105217 This copies the PropagateEquals logical optimization into ESQL, following the pattern established in #106499. I've copied the optimization rule into the ESQL version of OptimizerRules, and the tests into OpitmizerRulesTests, and changed the imports &c to point to the appropriate ESQL classes instead of their QL counterparts. I expect to have several more PRs following this pattern, for the remaining logical optimizations that touch the binary comparison logic. I'm intending to make separate PRs for each, in the interest of making them easier to review. | 22 March 2024, 16:48:00 UTC |
| cbc4182 | Fernando Briano | 22 March 2024, 15:25:38 UTC | Updates REST API test issue69009.yml (#106663) | 22 March 2024, 15:25:38 UTC |
| fb17da0 | yashdamani | 22 March 2024, 15:06:59 UTC | Don't trim highlight snippet if number_of_fragments is 0 Don't trim highlight snippet if number_of_fragments is 0 Closes #101803 | 22 March 2024, 15:06:59 UTC |
| 0eca8dd | Tim Rühsen | 22 March 2024, 14:53:39 UTC | [Profiling] Switch to OTEL cloud.provider, cloud.region, host.type (#106656) * [Profiling] Switch to OTEL cloud.provider, cloud.region, host.type * Remove wrong spaces from auto-format | 22 March 2024, 14:53:39 UTC |
| bdce52e | Tanguy Leroux | 22 March 2024, 14:33:32 UTC | Assert that FsBlobContainer.readBlob does not start reading after file length (#106668) This change adds an assertion in the FsBlobContainer.readBlob to ensure we are not reading after the last byte of the file. While this is legal and documented in the SeekableByteChannel.position() API, having this assertion in place would have caught on CI some regression introduced recently and only caught on deployments where S3 rejects reads starting after blob length. | 22 March 2024, 14:33:32 UTC |
| cc96a8b | Kostas Krikellas | 22 March 2024, 14:31:51 UTC | Add DownsampleMetrics (#106632) A single instance gets created by the `Downsample` plugin and injected to related actions. | 22 March 2024, 14:31:51 UTC |
| e92420d | Liam Thompson | 22 March 2024, 14:28:30 UTC | [DOCS] Update cross cluster search compatability matrix (#106677) | 22 March 2024, 14:28:30 UTC |
| faf7b7f | Martijn van Groningen | 22 March 2024, 13:50:01 UTC | Add more debugging to 25_id_generation/generates a consistent id test (#106667) Include the _ts_routing_hash field in the search response for more debugging. In order to add support for docvalue_fields feature for this field type, a custom DOC_VALUE_FORMAT is needed in order to decode to valid utf8 string. Relates #106550 | 22 March 2024, 13:50:01 UTC |
| 9a90770 | David Turner | 22 March 2024, 13:46:46 UTC | Move `XContent` -> `SnapshotInfo` parsing out of prod (#106669) The code to parse a `SnapshotInfo` object out of an `XContent` response body is only used in tests, so this commit moves it out of the production codebase and into the test framework. | 22 March 2024, 13:46:46 UTC |
| f02adac | Andrei Dan | 22 March 2024, 13:38:21 UTC | Fix testDataStreamLifecycleDownsampleRollingRestart (#106664) This removes a redundant thread creation when triggering a rolling restart as the method is already async and drops the check for cluster health as that might hit a node that's being shut down (the master node in particular) The test assertions still hold i.e. the successful downsampling of the source index | 22 March 2024, 13:38:21 UTC |
| bd6f3c7 | Benjamin Trent | 22 March 2024, 13:18:43 UTC | Test mute for issue #106647 (#106670) | 22 March 2024, 13:18:43 UTC |
| 391f010 | Luigi Dell'Aquila | 22 March 2024, 13:01:31 UTC | ES|QL: Fix usage of IN operator with TEXT fields (#106654) | 22 March 2024, 13:01:31 UTC |
| 46b5596 | David Turner | 22 March 2024, 12:47:47 UTC | Avoid double-sorting results in TransportGetSnapshotsAction (#106644) Today we call `sortSnapshots` twice, once for the per-repository results, and then again for the combined results. The first call has no limit or offset, and the sorting is made redundant by the second call, so only really serves to filter out snapshots which do not match the `?after` parameter. Meanwhile the `?after` parameter filtering is redundant in the second call. This commit separates these two steps to avoid the redundant sorting in the first step and the redundant filtering in the second. | 22 March 2024, 12:47:47 UTC |
| 16787c2 | Daniel Mitterdorfer | 22 March 2024, 12:14:07 UTC | [Profiling] Fix spurious test failures (#106660) With this commit we force-merge indices to workaround elastic/elasticsearch#106657. We also revert a test mute that was applied to the wrong test case so all integration tests are executed again. Relates #106309 Relates #106657 Closes #106308 | 22 March 2024, 12:14:07 UTC |
| f3f117a | Liam Thompson | 22 March 2024, 11:52:30 UTC | [DOCS] Update params for Update Connector Filtering API (#106662) | 22 March 2024, 11:52:30 UTC |
| d0c7da6 | Jim Ferenczi | 22 March 2024, 11:32:56 UTC | Improves Error Handling in Chunked Inference API (#106634) Previously, if one input failed within a batch call to the chunked inference API, the entire request was labeled as failed. This update alters this process, now allowing individual errors to be conveyed in the response. This empowers the caller to address failures specifically and understand the root cause, rather than receiving a vague error message like: 'Expected a chunked inference [chunked_text_embedding_service_results] received [error]'. | 22 March 2024, 11:32:56 UTC |
| c41df74 | Armin Braun | 22 March 2024, 11:31:16 UTC | Move more XContent parsers that are only used in tests to test codebase (#105801) Just like a couple times before, moving a couple more of the test only parsers to the test codebase to save code-size etc. | 22 March 2024, 11:31:16 UTC |
| a8188f8 | David Kyle | 22 March 2024, 11:12:45 UTC | [ML] Fix Array out of bounds exception in the XLM Roberta tokenizer (#106655) Increases the buffer size for the normalised form of the input unicode character. Certain characters can have surprisingly long normalised forms | 22 March 2024, 11:12:45 UTC |
| d8fc877 | Tanguy Leroux | 22 March 2024, 11:06:27 UTC | Fix S3RepositoryThirdPartyTests.testReadFromPositionLargerThanBlobLength (#106466) The test should use a random operation purpose that is not "Indices", otherwise S3RetryingInputStream retries up to Integer.MAX_VALUE times which causes the test suite to timeout. Also fixes the progress in the retries log messages. Closes #106457 | 22 March 2024, 11:06:27 UTC |
| 14ca58c | Dmitry Cherniachenko | 22 March 2024, 10:27:45 UTC | `Sets` utility class code cleanup (#105350) * `Sets` utility class code cleanup - newHashSetWithExpectedSize() returns HashSet, same as newHashSet() - haveEmptyIntersection() iterates a smaller set, same as intersection() - sortedDifference() accepts sets with Comparable values - replace Set.copyOf() in addToCopy() with Set.of() to avoid 1 extra copy of values | 22 March 2024, 10:27:45 UTC |
| 69ecdc6 | Nikolaj Volgushev | 22 March 2024, 10:26:45 UTC | Expose workflow restriction parsing (#106648) Need this one for custom role parsing upstream. | 22 March 2024, 10:26:45 UTC |
| 4647691 | Ievgen Degtiarenko | 22 March 2024, 10:00:39 UTC | Clarify the difference between ImmutableEntry and Map.entry (#106588) | 22 March 2024, 10:00:39 UTC |
| 23278a5 | David Turner | 22 March 2024, 09:50:18 UTC | Expand conceptual docs around `ActionListener` (#105405) Rephrases the docs about `ActionListener` in terms of continuations and control flow, rather than as injecting logic into lower layers. | 22 March 2024, 09:50:18 UTC |
| b5923eb | David Turner | 22 March 2024, 08:56:15 UTC | AwaitsFix for #106650 | 22 March 2024, 08:56:24 UTC |
| cc8fb4d | Alexander Spies | 22 March 2024, 08:42:29 UTC | ESQL: Re-enable logical dependency check (#105860) * Fix Enrich refs locally in the dependency check * Re-enable logical dependency check + test for it * Return server error on failed dependency check | 22 March 2024, 08:42:29 UTC |
| 1fe4946 | David Turner | 22 March 2024, 08:18:48 UTC | Fix testScheduledFixedDelayRejection (#106630) Not really necessary to allow the scheduled task to race against the blocks, and this race is a source of test flakiness. Fixed by imposing the blocks first. Closes #106618 | 22 March 2024, 08:18:48 UTC |
| 12e64a1 | Moritz Mack | 22 March 2024, 08:11:30 UTC | Use historical features for Yaml REST tests for cluster apis (#106587) | 22 March 2024, 08:11:30 UTC |
| a960bef | Lorenzo Dematté | 21 March 2024, 23:52:44 UTC | Adding `resolveConcreteIndexNames` to DataStreamsActionUtil (#106621) Extract method `resolveConcreteIndexNames` from `DataStreamsStatsTransportAction` to `DataStreamsActionUtil`, and move the whole class to the `server` module, so it can be shared and used across plugins. | 21 March 2024, 23:52:44 UTC |
| ff91ce0 | Athena Brown | 21 March 2024, 23:17:25 UTC | Update and pin gradle wrapper validation action (#106635) This commit updates the Gradle wrapper validation action to the latest release and pins it to that version by hash per our company best practices for GitHub Workflows & Actions. | 21 March 2024, 23:17:25 UTC |
| 2b67444 | Ryan Ernst | 21 March 2024, 22:02:26 UTC | Wait indefintely for http connections on shutdown by default (#106511) This commit changes the meaning of the http shutdown timeout when set to 0. Previously the shutdown would proceed immediately, killing active http connections. With this change 0 means no timeout is used, but active connections are still waited on, indefinitely. This new behavior is what was originally intended for the default value of the shutdown timeout setting. | 21 March 2024, 22:02:26 UTC |
| a400e3c | Nhat Nguyen | 21 March 2024, 21:32:49 UTC | Revert "Avoid using small inactive exchange timeout in breaker tests (#106394)" This reverts commit bce11c6c1c1c73a4b9a57b367cc6d379b9acac1b. | 21 March 2024, 21:32:49 UTC |
| e5b6040 | caichangheng | 21 March 2024, 18:31:50 UTC | Check preTags and postTags params for empty values (#106396) Check highlighter params of pre_tags and post_tags params for empty values. If empty throw IllegalArgumentException. Closes #69009 | 21 March 2024, 18:31:50 UTC |
| 3f8cc1e | Ryan Ernst | 21 March 2024, 18:19:12 UTC | Add test sourceset support to MRJAR plugin (#106571) Mrjar plugin supports main sourcesets automatically configured with newer java versions. This commit adds support for test sourcesets as well. The key difference between main and test is that test sourcesets also extend their main counterparts. | 21 March 2024, 18:19:12 UTC |
| edefbbc | Andrei Dan | 21 March 2024, 17:15:52 UTC | Install the stack templates only if DSL available (#106500) Data stream lifecycle has been made available only in 8.11 however we now have a stack template that makes use of it (`.kibana-reporting`). In an 8.10-8.14 mixed cluster environment we could cause a cluster state corruption if we install the `kibana-reporting` (with `lifecycle:{}`) template before all nodes _know_ of data stream lifecycle (8.10 does not). This changes the `StackTemplateRegistry` to wait for all nodes in the cluster to be aware of data stream lifecycle before installing the index templates. | 21 March 2024, 17:15:52 UTC |
| 47b6eb3 | Moritz Mack | 21 March 2024, 17:12:22 UTC | Re-mute SimpleThreadPoolIT.testThreadPoolMetrics (#106623) Mute test until remaining races are addressed. (related to #104652) | 21 March 2024, 17:12:22 UTC |
| 7c46c73 | Nik Everett | 21 March 2024, 17:01:15 UTC | ESQL: Generate docs for ceil (#106616) This replaces the hand maintained docs for `CEIL` and with the docs generated by the tests. There shouldn't be any diff in the generated docs. | 21 March 2024, 17:01:15 UTC |
| 25d586c | Mary Gouseti | 21 March 2024, 16:59:27 UTC | Make testXContentSerializationWithRolloverAndEffectiveRetention take into consideration a disabled lifecycle (#106612) | 21 March 2024, 16:59:27 UTC |
| 9c8c319 | Luigi Dell'Aquila | 21 March 2024, 16:53:29 UTC | Refactor ESQL HeapAttackIT to run on Serverless (#106586) | 21 March 2024, 16:53:29 UTC |
| a41e11f | Mark Vieira | 21 March 2024, 16:52:12 UTC | Improve stability of build complete archive creation (#106578) This is an attempt to fix some errors creating the diagnostic archive at the end of builds in CI. We've seen errors where we attempt to write more bytes than expected into the archive. The assumption here is that we are bundling files that are possibly still being written to. This _shouldn't_ happen, but it's useful to have these diagnostic bundles in either case. Closes https://github.com/elastic/elasticsearch/issues/106576 | 21 March 2024, 16:52:12 UTC |
| fa00e61 | Nik Everett | 21 March 2024, 16:52:04 UTC | ESQL: Values aggregation function (#106065) This creates the `VALUES` aggregation function which buffers all field values it receives and emits them as a multivalued field. It can use a significant amount of memory and will circuit break if it uses too much memory, but it's really useful for putting together self-join-like behavior. It sort of functions as a stop-gap measure until we have more self-join style things. In the future we'll have spill-to-disk for aggregations and, likely, some kind of self-join command for aggregations at least so this will be able to grow beyond memory. But for now, memory it is. Example: ``` FROM employees | EVAL first_letter = SUBSTRING(first_name, 0, 1) | STATS first_name=VALUES(first_name) BY first_letter | SORT first_letter ; first_name:keyword | first_letter:keyword [Anneke, Alejandro, Anoosh, Amabile, Arumugam] | A [Bezalel, Berni, Bojan, Basil, Brendon, Berhard, Breannda] | B [Chirstian, Cristinel, Claudi, Charlene] | C [Duangkaew, Divier, Domenick, Danel] | D ``` I made this work for everything but `geo_point` and `cartesian_point` because I'm not 100% sure how to integrate with those. We can grab those in a follow up. Closes #103600 | 21 March 2024, 16:52:04 UTC |
| 93a21e1 | David Turner | 21 March 2024, 16:30:18 UTC | AwaitsFix for #106618 | 21 March 2024, 16:38:09 UTC |
| a8586f3 | David Turner | 21 March 2024, 16:31:10 UTC | Document lack of logging BwC (#104459) Our guarantees around backwards compatibility and breaking changes do not apply to the Elasticsearch application logs. This commit adds a note to the docs about this. | 21 March 2024, 16:31:10 UTC |
| 76d1974 | Rene Groeschke | 21 March 2024, 16:29:38 UTC | Add gradle plugin for publishing docker based test fixtures (#106229) To make those docker images available for our testcontainer based fixtures we need to publish them on a regular basis to ensure the source and the images are in sync. This adds some convenience plugin to take care of publishing our docker test fixtures | 21 March 2024, 16:29:38 UTC |
| 521e06d | Ryan Ernst | 21 March 2024, 16:00:53 UTC | Remove node shutdown test check on snapshot build (#106518) Node shutdown is no longer guarded by a feature flag. This commit removes the test assumptions that only allowed some tests to run on snapshots which enable feature flags. | 21 March 2024, 16:00:53 UTC |
| de7e898 | István Zoltán Szabó | 21 March 2024, 15:41:31 UTC | [DOCS] Amends important note on delayed data detection. (#106610) | 21 March 2024, 15:41:31 UTC |
| bce11c6 | Nhat Nguyen | 21 March 2024, 15:24:32 UTC | Avoid using small inactive exchange timeout in breaker tests (#106394) The tests failed because we set an inactive exchange timeout too short, causing a timeout error instead of triggering the expected CircuitBreakingException. Closes #105681 Closes #105543 | 21 March 2024, 15:24:32 UTC |
| 3489906 | Nik Everett | 21 March 2024, 14:51:35 UTC | ESQL: Generate a few more docs (#106577) And improve the error message on csv test failures. | 21 March 2024, 14:51:35 UTC |
| 49d40c3 | Dianna Hohensee | 21 March 2024, 14:48:47 UTC | Complete placeholder AtomicRegister Coordinator tests (#106512) Porting over Coordinator tests to Stateless/AtomisRegisterCoordinatorTests | 21 March 2024, 14:48:47 UTC |
| f617694 | Mark Tozzi | 21 March 2024, 14:33:12 UTC | [ESQL] copy ql disjunction rule into esql (#106499) Similar to #105711, this migrates a QL optimization that depends on specific class types into ESQL to refer to the ESQL version of those classes. This is necessary for #105217, which changes the class hierarchy of the binary comparisons. I migrated the existing rules tests, and added new tests to the logical plan optimizer tests to cover this rule. | 21 March 2024, 14:33:12 UTC |
| dcc3d83 | Lisa Cawley | 21 March 2024, 14:20:56 UTC | [DOCS] Unhealthy transform rule check (#106521) | 21 March 2024, 14:20:56 UTC |
| ad47382 | Moritz Mack | 21 March 2024, 14:06:54 UTC | Wait in SimpleThreadPoolIT.testThreadPoolMetrics for threads to complete (#106580) Search threads might still be active even after returning a response. Waiting for all threads to be completed prevents a race between the collection of threadpool stats and measurements. | 21 March 2024, 14:06:54 UTC |
| 11a4e40 | István Zoltán Szabó | 21 March 2024, 13:57:10 UTC | [DOCS] Adjusts PUT inference API docs examples (#106604) Co-authored-by: David Kyle <david.kyle@elastic.co> | 21 March 2024, 13:57:10 UTC |
| dff4fd4 | Rassyan | 21 March 2024, 13:50:14 UTC | Log AffixSetting update when using addAffixMapUpdateConsumer (#97072) | 21 March 2024, 13:50:14 UTC |
| 26f7269 | Nik Everett | 21 March 2024, 13:46:23 UTC | ESQL: Escape regexes in tests (#106603) When we're generating raw strings to match we want to make sure it just matches exactly the string - so we should escape it. If we don't we'll get weird failures from syntax exceptions. Closes #106533 Closes #106538 | 21 March 2024, 13:46:23 UTC |
| e92a7a7 | István Zoltán Szabó | 21 March 2024, 13:33:10 UTC | [DOCS] Adds note to inference tutorial about similarity (#106567) | 21 March 2024, 13:33:10 UTC |
| 999dcb8 | Salvatore Campagna | 21 March 2024, 12:35:17 UTC | Unable to retrieve multiple stored field values (#106575) The issue happens when we try to use multiple stored fields through the FetchFieldsPhase, which we do when using `_fields` since we have a single shared instance of SingleFieldsVisitor per field and document and use a shared `currentValues` array. | 21 March 2024, 12:35:17 UTC |
| ffec5f6 | Martijn van Groningen | 21 March 2024, 12:17:19 UTC | Fix synthesizing id bug when reading routing value (#106600) Take `routingHashBytes` offset and length into account when decoding the routing hash as part of synthesizing the _id. | 21 March 2024, 12:17:19 UTC |
| e1e14e4 | Yang Wang | 21 March 2024, 11:29:20 UTC | Finer grained blobStore stats report in stateless (#106581) When stateless is enabled, the blobStore stats report keep the information about operation purposes by using combined keys of operation purpose and operation. Resolves: ES-7967 | 21 March 2024, 11:29:20 UTC |
| 8a7697b | Martijn van Groningen | 21 March 2024, 10:37:34 UTC | Improve short-circuiting downsample execution (#106563) This is relevant in the case multiple downsample api invocations have been executed for the same source index, target index and fixed interval. Whether the target index is ready, is now also checked just before starting the downsample persistent tasks. Relates to #106403 | 21 March 2024, 10:37:34 UTC |
| 32dbc28 | István Zoltán Szabó | 21 March 2024, 10:32:57 UTC | [DOCS] Adds disclaimer to semantic search tutorials (#106590) | 21 March 2024, 10:32:57 UTC |
| 992e9b5 | Tim Rühsen | 21 March 2024, 10:14:18 UTC | [Profiling] Switch to OTEL field host.arch (#106585) | 21 March 2024, 10:14:18 UTC |
| 2120683 | Przemysław Witek | 21 March 2024, 09:58:29 UTC | [Transform] Fix `_reset` API when called with `force=true` on a failed transform (#106574) | 21 March 2024, 09:58:29 UTC |
| c91409c | David Turner | 21 March 2024, 09:14:13 UTC | Improve threading in `TransportGetSnapshotsAction` (#106527) There's a comment about intentionally avoiding using the `SNAPSHOT_META` pool in `TransportGetSnapshotsAction`, but in fact we implicitly fork onto the `SNAPSHOT_META` pool in a few spots and end up doing much of the work there anyway. This commit fixes things to use the `MANAGEMENT` pool for all this work as originally intended. | 21 March 2024, 09:14:13 UTC |
| 4374a38 | Moritz Mack | 21 March 2024, 08:39:22 UTC | Use historical features for Yaml REST tests for bulk api (#106402) | 21 March 2024, 08:39:22 UTC |
| 19cffde | Kostas Krikellas | 21 March 2024, 06:00:37 UTC | Support non-keyword dimensions as routing fields in TSDB (#105501) * Fix test failure https://gradle-enterprise.elastic.co/s/icg66i6mwnjoi * Fix test failure https://gradle-enterprise.elastic.co/s/icg66i6mwnjoi * Nest pass-through objects within objects * Support numeric fields as pass-through dimensions * Update docs/changelog/105073.yaml * fix tests * fix tests * refactor dimension property * fix yaml * add numeric to routing builder * fix violation * fix yaml versions * handle all numeric values in dimension fields * use shardId in TSDB doc id * spotless fix * test fixes * index version guard * test fix * revert shardId changes * add routing id field mapper * Update docs/changelog/105501.yaml * fix shrunk test * use SortedDocValuesField * fix tests * delete unused yaml * submit test fixes * sync * restrict id override to latest version * update painless path * test fix * check for invalid id * fix painless tests * restore bwc routing changes * add unittest * refactor passthrough test * address comments * remove redundant variable * remove assert * fix id test * fix painless * skip storing routing field * skip generating the id when available * skip assert on routing in TranslogWriter * skip id mismatch assert * changelog update * revert redundant changes * remove supportsDimension * spotless * rename var for forcing dimension * small refactor * small refactor | 21 March 2024, 06:00:37 UTC |
| c4e9646 | Moritz Mack | 21 March 2024, 05:17:24 UTC | Re-enable the RemoteClusterSecurityBwcRestIT (#106570) This reverts commit 4ef0a8f1068f9701c087c54cdbd23f49742b7e7f / #106534 | 21 March 2024, 05:17:24 UTC |
| a130537 | Nik Everett | 20 March 2024, 20:32:12 UTC | ESQL: Use generated docs for abs and acos (#106510) ESQL: Use generated docs for abs and acos | 20 March 2024, 20:32:12 UTC |
| 2988799 | Mary Gouseti | 20 March 2024, 18:27:08 UTC | [DSL Global Retention] Use data stream global retention metadata (#106221) | 20 March 2024, 18:27:08 UTC |
| 52b311d | Martijn van Groningen | 20 March 2024, 18:19:24 UTC | Rename index names in 25_id_generation.yml file (#106558) Rename index names in yaml test to more easily find related log lines in es server log. Many yaml tests are being executed in the same qa cluster, this should help finding relevant logs for failures in this yaml file. Relates to #106550 | 20 March 2024, 18:19:24 UTC |
| 2523ed9 | Mark Vieira | 20 March 2024, 16:04:52 UTC | Validate that test cluster BWC nodes use the default distribution (#106559) We have instances where BWC tests configure old ES version nodes with the integTest distribution. This isn't a valid configuration, and while we in reality resolve the default distribution artifact, we have other configuration logic that behaves differently based on whether the integTest distro was _requested_. Specifically, what to set ES_JAVA_HOME to. This bug resulted in us attempting to run old nodes using the current bundled JDK version, which may be incompatible with that older version of Elasticsearch. Closes #104858 | 20 March 2024, 16:04:52 UTC |
| 43de7a4 | Ignacio Vera | 20 March 2024, 15:57:17 UTC | Introduce a generic HashMap backed by big arrays (#106418) this commit introduces a PagedHashMap which is equivalent of a Java HashMap but backed by big arrays. | 20 March 2024, 15:57:17 UTC |
| 58c9233 | Fang Xing | 20 March 2024, 15:45:54 UTC | [ES|QL] Centralize Datetime String Conversion (#106388) * centralize datetime string conversion | 20 March 2024, 15:45:54 UTC |
| 19cf318 | Ignacio Vera | 20 March 2024, 15:38:38 UTC | Use LogDocMergePolicy in GeoPointScriptFieldDistanceFeatureQueryTests#testMatches (#106557) | 20 March 2024, 15:38:38 UTC |
| b24e0d0 | Craig Taverner | 20 March 2024, 15:36:00 UTC | Fix 60_enrich/Enrich on keyword (fixes 106507) (#106513) * Mute failing test * Remove gradle mute, and change comment to indicate this is a fix It appears the difference that caused the problem is that the query now returns a field that used to be a keyword, but is now a text field, provided by the enrich lookup. So even though the enrich is still based on a keyword, this query now accesses code paths it did not before, and those have issues before 8.13.0 related to reading text from source. | 20 March 2024, 15:36:00 UTC |
| a368b48 | David Turner | 20 March 2024, 14:43:22 UTC | Release `TranslogSnapshot` buffer after iteration (#106398) Closes #106390 | 20 March 2024, 14:43:22 UTC |
| 7517424 | David Turner | 20 March 2024, 14:22:30 UTC | AwaitsFix for #106554 | 20 March 2024, 14:22:38 UTC |
| 5d05d81 | Fang Xing | 20 March 2024, 14:15:49 UTC | [ES|QL] Remove variadic functions' optional args from the output of show functions (#106454) * remove optional args from the output of show functions optionalArgs, argNames and argTypes for variadic functions * consistent names for arguments | 20 March 2024, 14:15:49 UTC |
| 09b6072 | István Zoltán Szabó | 20 March 2024, 14:07:41 UTC | [DOCS] Changes Cohere inference examples in tutorial and API docs (#106524) | 20 March 2024, 14:07:41 UTC |
| 61e8134 | Moritz Mack | 20 March 2024, 13:58:15 UTC | Unmute SimpleThreadPoolIT.testThreadPoolMetrics with additional logging (#106522) Add additional logging and more vocal assertions to better understand test failures (related to #104652) | 20 March 2024, 13:58:15 UTC |
| 2e8482a | David Turner | 20 March 2024, 13:47:14 UTC | Integrate threadpool scheduling with `AbstractRunnable` (#106542) Today `ThreadPool#scheduleWithFixedDelay` does not interact as expected with `AbstractRunnable`: if the task fails or is rejected then this isn't passed back to the relevant callback, and the task cannot specify that it should be force-executed. This commit fixes that. | 20 March 2024, 13:47:14 UTC |
| cf5fbfc | David Turner | 20 March 2024, 13:28:59 UTC | Force execution of `SearchService.Reaper` (#106544) If the search threadpool fills up then we may reject execution of `SearchService.Reaper` which means it stops retrying. We must instead force its execution so that it keeps on going. With #106542, closes #106543 | 20 March 2024, 13:28:59 UTC |
| 82221ff | Aurélien FOUCRET | 20 March 2024, 12:35:05 UTC | Ensure ILM policy is installed before starting the tests. (#106523) | 20 March 2024, 12:35:05 UTC |
| 9315f80 | Henning Andersen | 20 March 2024, 12:20:54 UTC | Enhance search tier GC options (#106526) For small nodes, we need a bit more wiggle room for new size and concurrent GC threads in order to stay below real memory circuit breaker limits on dedicated search nodes. ES-8087 | 20 March 2024, 12:20:54 UTC |
| 2087b65 | David Kyle | 20 March 2024, 11:15:21 UTC | [ML] Create inference_user and inference_admin roles (#106371) Defines new inference_user and inference_admin roles with the related cluster privileges manage_inference and monitor_inference. inference_user can list the models and preform inference, inference_admin can do the same plus create and delete models | 20 March 2024, 11:15:21 UTC |
| 0d6b368 | David Roberts | 20 March 2024, 11:03:21 UTC | [ML] Fork trained model stats work to thread pool threads (#99439) The steps that assemble the trained model stats response largely ran on the transport worker thread. Even though none of the steps look massively expensive, it's likely one of them is, because we've observed client lockups (Kibana) while trained model stats is running. This change moves the steps into the ML utility thread pool to avoid any risk of blocking the transport thread. --------- Co-authored-by: David Turner <david.turner@elastic.co> | 20 March 2024, 11:03:21 UTC |
| 504ef7c | David Turner | 20 March 2024, 10:55:56 UTC | Clarify threading in `SubscribableListener` (#106532) The docs about threading here are a little opaque to consumers, who typically just want to ensure that the listener is completed on a particular threadpool. This commit adds clearer docs about achieving that goal. | 20 March 2024, 10:55:56 UTC |
| 4ef0a8f | Simon Cooper | 20 March 2024, 10:40:35 UTC | Revert "Re-enable the RemoteClusterSecurityBwcRestIT (#106052)" (#106534) This reverts commit 9b0f1cea8ceb603e1990bef99b5f7e895d7f617d. | 20 March 2024, 10:40:35 UTC |
| fdd6d89 | Navarone Feekery | 20 March 2024, 10:30:18 UTC | [Connectors API] Add missing `_api_key_id` docs (#106469) | 20 March 2024, 10:30:18 UTC |
| 5afc59b | István Zoltán Szabó | 20 March 2024, 10:05:52 UTC | [DOCS] Creates a semantic_text field type docs page. (#106528) | 20 March 2024, 10:05:52 UTC |
| 9980eac | David Turner | 20 March 2024, 09:59:47 UTC | AwaitsFix for #106533 | 20 March 2024, 09:59:55 UTC |
| 430758d | Nikolaj Volgushev | 20 March 2024, 09:28:47 UTC | Test tweaks for compatible security tests (#106464) Test tweaks for serverless: * Valid application name in API key tests * Move from `cluster.health` to `info` call in roles test (the call is just used to check that a user with a cluster privilege is indeed able to execute the test) Closes: ES-7987 | 20 March 2024, 09:28:47 UTC |
| d01adff | Ioana Tagirta | 20 March 2024, 09:03:04 UTC | Add links to text_expansion in ELSER tutorial (#106490) * Add links to text_expansion in ELSER tutorial * Apply suggestions from code review Co-authored-by: Liam Thompson <32779855+leemthompo@users.noreply.github.com> --------- Co-authored-by: Liam Thompson <32779855+leemthompo@users.noreply.github.com> | 20 March 2024, 09:03:04 UTC |
| 32bc182 | Chris Hegarty | 20 March 2024, 08:50:16 UTC | ESQL: expose transport action to improve usage (#106463) This commits exposes the query transport action to improve usage. While one can perform all operations prior to this change, it has been suggested that adding the action would improve the symmetry of the API by allowing e.g. client().execute(builder.action(), builder.request()).actionGet(30, SECONDS); | 20 March 2024, 08:50:16 UTC |
| 6f607e4 | Lorenzo Dematté | 20 March 2024, 08:27:31 UTC | Use cluster features for ASYNC ESQL tests (#104466) + Add esql as rest test dependency for ml/native-multi-node-tests to work around the mixed testClusters/TestCluster nodes (so all have the esql plugin installed) | 20 March 2024, 08:27:31 UTC |
| c30999b | Yang Wang | 20 March 2024, 08:05:43 UTC | [Test] Ranged read should read non-empty content (#106000) Empty read is [short-circuited](https://github.com/elastic/elasticsearch/blob/e8039b9ecb2451752ac5377c44a6a0c662087a9f/modules/repository-s3/src/main/java/org/elasticsearch/repositories/s3/S3BlobContainer.java#L115-L116) without going to the blob store. In order to test s3 blob store, ranged read should read at least one byte. This PR ensures that. Resolves: #105958 | 20 March 2024, 08:05:43 UTC |
| dbb7847 | Albert Zaharovits | 20 March 2024, 08:05:02 UTC | Expose lookup of realm domain config by realm id (#106424) The scope here is to expose a method (Realms#getRealmRef) that can be used to retrieve the realm domain assignments for any realm id. | 20 March 2024, 08:05:02 UTC |
| d5565b6 | Athena Brown | 19 March 2024, 23:08:41 UTC | Fix typo in OIDC docs (#106207) (#106517) Add missing _to_ in sentence (cherry picked from commit 40a9155b3b53925b335530da4e076e39e66ee8f7) Co-authored-by: Aaron Hanusa <aaron.hanusa@gmail.com> | 19 March 2024, 23:08:41 UTC |
