| ff8f75c | Armin Braun | 05 April 2024, 01:25:25 UTC | Remove unused ShrinkAction (#107127) ShrinkAction isn't used anymore, we can remove it and its trivial test coverage. | 05 April 2024, 01:25:25 UTC |
| b3b4214 | Mark Vieira | 04 April 2024, 20:25:11 UTC | Skip multi-release source sets in idea project import by default (#107123) There is an existing IntelliJ bug that prevents doing a full project build when source sets for multi-release jars are present. This changes the project import behavior so that these source sets are ignored by default and can be explicitly enabled by adding `org.gradle.mrjar.idea.enabled=true` to your `~/.gradle/gradle.properties` file should you need to actively work on that code. | 04 April 2024, 20:25:11 UTC |
| 3486a08 | Armin Braun | 04 April 2024, 19:01:06 UTC | Remove unused methods from SearchContext (#107111) A couple methods can be removed here, they're only used on sub-classes. That also cleans up some unsupport operation exception code that is never hit. | 04 April 2024, 19:01:06 UTC |
| a5e7525 | Costin Leau | 04 April 2024, 18:22:28 UTC | ESQL: Disable flaky test | 04 April 2024, 18:22:37 UTC |
| c2bd3e4 | Costin Leau | 04 April 2024, 17:59:24 UTC | ESQL: Fix flaky test in LogicalPlanOptimizerTests | 04 April 2024, 17:59:24 UTC |
| 7483844 | Costin Leau | 04 April 2024, 17:12:41 UTC | ESQL: Allow grouping key inside stats expressions (#106579) Similar to aggs, allow grouping keys to used inside STATS expressions by introducing a synthetic eval, e.g.: STATS a = x + count(*) BY x becomes STATS c = count(*) BY x | EVAL a = x + c | KEEP a, x To better handle overriding aliases, introduce EsqlAggregate which keeps the declared structure intact during analysis and verification while merging the output. The deduplication happens now in the optimization phase. Fix small bug that caused replacement of expressions inside aggregations to be skipped despite being applied Improved Verifier to not repeat error messages in case for Aggregates Removed verification heuristics for missing columns as functions as it was too broad | 04 April 2024, 17:12:41 UTC |
| 0cc19f3 | James Baiera | 04 April 2024, 16:18:29 UTC | Extract failure store specific settings to the failure store (#107063) Moves the failure store specific settings logic to the same place that we manage their mappings. | 04 April 2024, 16:18:29 UTC |
| 75f5487 | Fang Xing | 04 April 2024, 15:34:35 UTC | fix substring type resolution (#107109) | 04 April 2024, 15:34:35 UTC |
| 6c986cb | Volodymyr Krasnikov | 04 April 2024, 15:20:08 UTC | Add primary and coordination operation rejection ratio metrics (#106978) Introduce two new metrics `es.indexing.primary_operations.rejections.ratio` and `es.indexing.coordinating_operations.rejections.ratio` They are needed as a second signal (along with rejections rate) to determine if it is time to alert about 429s error during _bulk indexing + add IT test to ensure that indexing metrics are published | 04 April 2024, 15:20:08 UTC |
| 01efbbf | Niels Bauman | 04 April 2024, 14:51:02 UTC | Remove AwaitsFix of fixed test (#107104) This `@AwaitsFixed` linked to https://github.com/elastic/elasticsearch/issues/102813, which was marked as a duplicate of https://github.com/elastic/elasticsearch/issues/102337, which in turn was fixed by https://github.com/elastic/elasticsearch/pull/102724. | 04 April 2024, 14:51:02 UTC |
| 08ad143 | Jake Landis | 04 April 2024, 14:46:33 UTC | Support swapping the user context to the secondary auth user for named actions (#106613) This commit adds the ability to define specific actions that are required to be executed as the secondary authenticated user. When actions are defined as secondary auth actions, then the secondary authentication headers (and subsequent authentication) are required to call those actions and the authorization is based on the secondary user. The SPI hook to define which actions are required is intended for internal only use only. | 04 April 2024, 14:46:33 UTC |
| 46ec636 | Daniel Mitterdorfer | 04 April 2024, 14:45:06 UTC | [Profiling] Add TopN Functions API (#106860) With this commit we add a new API to the Universal Profiling plugin that allows to gather a list of functions with the most observed samples (TopN functions). --------- Co-authored-by: Joseph Crail <joseph.crail@elastic.co> | 04 April 2024, 14:45:06 UTC |
| 8a1df9b | shainaraskas | 04 April 2024, 14:44:14 UTC | [DOCS] fix time zone logic example (#106962) * [DOCS] fix time zone logic example * specify standard time * goodbye e.g. | 04 April 2024, 14:44:14 UTC |
| edc9e67 | Alexander Spies | 04 April 2024, 14:28:07 UTC | ESQL: Fix fully pruned aggregates (#106673) Fix a bug where PruneColumns would sometimes completely replace an Aggregate, producing the wrong number of rows as a result. | 04 April 2024, 14:28:07 UTC |
| abfb0ae | Alexander Spies | 04 April 2024, 14:24:33 UTC | ESQL: Fix treating all fields as MV in COUNT pushdown (#106720) Fix a mistake in #106690 that accidentally prevented COUNT(field) from being pushed down in case field is single-valued. Add test to avoid future regressions. | 04 April 2024, 14:24:33 UTC |
| da28e76 | Andrei Dan | 04 April 2024, 14:20:14 UTC | Default `data_streams.auto_sharding.excludes` to NO exclusions (`[]`) (#107091) | 04 April 2024, 14:20:14 UTC |
| 87a9953 | Volodymyr Krasnikov | 04 April 2024, 14:11:47 UTC | Unmute testUniDirectionalIndexFollowing (#106886) | 04 April 2024, 14:11:47 UTC |
| 29a3256 | István Zoltán Szabó | 04 April 2024, 13:42:03 UTC | [DOCS] Makes Inference APIs main page more informative. (#107100) | 04 April 2024, 13:42:03 UTC |
| f30b79e | Jedr Blaszyk | 04 April 2024, 13:19:18 UTC | [Connector API] Support updating single schedule type (full, incremental or access_control) (#107090) | 04 April 2024, 13:19:18 UTC |
| 43d0ef9 | David Turner | 04 April 2024, 13:05:28 UTC | AwaitsFix for #107043 | 04 April 2024, 13:05:37 UTC |
| 90351ef | David Turner | 04 April 2024, 13:04:14 UTC | Capture hot threads during slow cluster state application (#107087) Today we emit a warning when applying the cluster state takes more than 30s by default. Experience shows that it's almost always one task that exceeds the limit, and usually that task is the `IndicesClusterStateService` (see #89821). This commit adds a `DEBUG` logger that will capture a thread dump if a task is running for longer than a configurable duration so we can get more insight into the cause of the slow execution. | 04 April 2024, 13:04:14 UTC |
| e33fd1b | Jedr Blaszyk | 04 April 2024, 12:06:12 UTC | [Connector API] Support numeric for configuration select option value type (#107059) | 04 April 2024, 12:06:12 UTC |
| d6582cf | Luca Cavanna | 04 April 2024, 10:42:23 UTC | Address concurrency issue in top hits aggregation (#106990) Top hits aggregation runs the fetch phase concurrently when the query phase is executed across multiple slices. This is problematic as the fetch phase does not support concurrent execution yet. The core of the issue is that the search execution context is shared across slices, which call setLookupProviders against it concurrently, setting each time different instances of preloaded source and field lookup providers. This makes us cross streams between slices, and hit lucene assertions that ensure that stored fields loaded from a certain thread are not read from a different thread. We have not hit this before because the problem revolves around SearchLookup which is used by runtime fields. TopHitsIT is the main test we have for top hits agg, but it uses a mock script engine which bypasses painless and SearchLookup. | 04 April 2024, 10:42:23 UTC |
| a32512f | Jedr Blaszyk | 04 April 2024, 08:51:15 UTC | Connector API: Followup on #106060 (#107058) | 04 April 2024, 08:51:15 UTC |
| 7b25421 | Ioana Tagirta | 04 April 2024, 07:48:35 UTC | Add ES|QL signum function (#106866) * Add ES|QL signum function * Update docs/changelog/106866.yaml * Skip csv tests for versions older than 8.14 * Reference layout docs file and fix instructions for adding functions * Break csv specs by param type * More tests | 04 April 2024, 07:48:35 UTC |
| 996a164 | Niels Bauman | 04 April 2024, 07:47:14 UTC | Set visibility of failure_store param of Rollover API to feature_flag (#107061) | 04 April 2024, 07:47:14 UTC |
| daa9006 | David Turner | 04 April 2024, 07:17:14 UTC | Apply snapshot `?after` filter inline (#107003) In `TransportGetSnapshotsAction` today we build a list of all candidate snapshots and then copy them into another list to apply the `?after` filter. With this commit we construct the final filtered list directly. | 04 April 2024, 07:17:14 UTC |
| 9a2f8a8 | David Turner | 04 April 2024, 06:37:13 UTC | Add remote cluster network troubleshooting docs (#107072) Spells out in a little more detail our expectations for remote cluster connections, including an example log message when the network is unreliable and some suggestions for how to troubleshoot further. | 04 April 2024, 06:37:13 UTC |
| 72a8248 | Daniel Mitterdorfer | 04 April 2024, 05:05:18 UTC | [Profiling] Use default task cancellation check (#107037) With this commit we remove our custom implementation of whether a task has been cancelled and instead use the standard implementation that is already provided by the task API. | 04 April 2024, 05:05:18 UTC |
| a37debd | Jake Landis | 03 April 2024, 21:40:41 UTC | Ensure getUser() is the logical user, not API key creator for RCS 2.0 (#107023) This commit changes SecurityContext#getUser() to provide the original user that initiated the call when run across clusters for RCS 2.0. Before this change the getUser() would provide the RCS 2.0 API key creator as the current user. | 03 April 2024, 21:40:41 UTC |
| 147f5a0 | Alexander Spies | 03 April 2024, 19:23:51 UTC | ESQL: Introduce language versioning to REST API (#106824) For the _query endpoint, add a parameter for the ESQL language version to the JSON payload. For now, it is optional and is only validated with no further action. | 03 April 2024, 19:23:51 UTC |
| f49ead7 | Max Hniebergall | 03 April 2024, 18:48:22 UTC | [ML] Add Cohere rerank to _inference service (#106378) | 03 April 2024, 18:48:22 UTC |
| 89563c2 | Andrei Dan | 03 April 2024, 17:27:13 UTC | Fix auto sharding recommending 0 shards for 0.0 workload (#107064) | 03 April 2024, 17:27:13 UTC |
| 8cab439 | Lee Hinman | 03 April 2024, 17:15:48 UTC | Mark enrich stats API as internal instead of public (#107022) This was previously publically available, however it exposes node counts and IDs. After discussion we've decided it's unnecessary on Serverless. Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com> | 03 April 2024, 17:15:48 UTC |
| 42df8fe | Tim Grein | 03 April 2024, 17:02:24 UTC | [Inference API] Make completion task available in inference processor (#107060) | 03 April 2024, 17:02:24 UTC |
| d88836b | David Kyle | 03 April 2024, 16:19:23 UTC | [ML] Add missing TokenizationConfigUpdate named writable to registry (#107056) | 03 April 2024, 16:19:23 UTC |
| 2f33b01 | James Baiera | 03 April 2024, 16:08:48 UTC | Split the mappings for failure stores out of the index template service. (#107025) This will help us keep failure store specific configurations in one place, and hopefully make it easier to evolve the schema in the future. | 03 April 2024, 16:08:48 UTC |
| 245d69d | Ash | 03 April 2024, 15:53:38 UTC | [Security Solution] Add `read` permission for third party agent indices for `kibana_system` (#107046) * add read permission for third party agent indices Allow `read` privilege for `kibana_system` role on `logs-sentinel_one*` and `logs-crowdstrike*` index patterns closes elastic/security-team/issues/9046 * Update docs/changelog/107046.yaml * review changes | 03 April 2024, 15:53:38 UTC |
| 36bcb6b | Albert Zaharovits | 03 April 2024, 15:33:14 UTC | Query API Keys support for both `aggs` and `aggregations` keywords (#107054) The Query API Key Information endpoint supports aggs since #104895. But some lang clients actually use the `aggregations` keyword in requests, as the preferred synonym to `aggs`. This PR adds support for the `aggregations` request keyword as a synonym for the existing `aggs` term. Closes #106839 | 03 April 2024, 15:33:14 UTC |
| 755226d | Kostas Krikellas | 03 April 2024, 15:09:17 UTC | [TEST] Add full cluster restart test for downsampling (#107053) This is a copy of the test added for rolling upgrade in #107036 | 03 April 2024, 15:09:17 UTC |
| 6e36089 | Joe Reuter | 03 April 2024, 15:04:24 UTC | Add non-indexed fields to ecs templates (#106714) * add non-indexed fields to ecs templates * update * Update docs/changelog/106714.yaml * Adjusting ECS tests to verify proper index and doc_values settings * review comment --------- Co-authored-by: eyalkoren <41850454+eyalkoren@users.noreply.github.com> | 03 April 2024, 15:04:24 UTC |
| f9ae6db | Bogdan Pintea | 03 April 2024, 14:23:36 UTC | ESQL: Add docs for the OPTIONS directive (#107013) This adds the docs for the newly added `OPTIONS` directive to `FROM`. | 03 April 2024, 14:23:36 UTC |
| 377c7e5 | Martijn van Groningen | 03 April 2024, 14:01:50 UTC | Implement time series grouping via HashAggregationOperator using a specialized block hash implementation (#106127) This change adds new operator factory that performs grouping by `_tsid` and `@timestamp` field. The new operator factory delegates to the `HashAggregationOperator` do the grouping, but does make use of a specialized block hash (which makes a few assumptions about the group fields). This is an initial implementation that is not optimize for 'local' grouping. The new time series grouping operator factory isn't usable from API and is only usable from tests. | 03 April 2024, 14:01:50 UTC |
| 333c6a6 | Salvatore Campagna | 03 April 2024, 13:58:57 UTC | Replace UnsupportedOperationException with IllegalArgumentException (#107038) This makes sure the Rest status we return is a 4xx instead of a 5xx, which means the error does not map anymore to "Service Unavailable". | 03 April 2024, 13:58:57 UTC |
| 206a0b7 | shainaraskas | 03 April 2024, 13:54:53 UTC | [DOCS] Remove obsolete accounting circuit breakers (#107015) | 03 April 2024, 13:54:53 UTC |
| ea9e6a9 | shainaraskas | 03 April 2024, 13:48:31 UTC | Clarify closed index setting warning (#106888) | 03 April 2024, 13:48:31 UTC |
| b97e2d6 | Nik Everett | 03 April 2024, 13:46:32 UTC | ESQL: Fixup docs for LOG and LOG10 (#106963) This merges all of the hand written docs for `LOG` and `LOG10` into the annotations which updates the `META FUNCTIONS` - now it'll always be the same as the docs. This also deletes the hand maintained docs and let's the documentation generation process rebuild it. | 03 April 2024, 13:46:32 UTC |
| 8dff321 | Luca Cavanna | 03 April 2024, 13:46:03 UTC | Remove allowedFields argument from SearchExecutionContext constructor (#107021) allowedFields is a mutable instance member, that only ever gets set via its setter. It does not need to be provided via constructor. Removing a needless argument seems important given the number of other arguments already present. | 03 April 2024, 13:46:03 UTC |
| a3a60b5 | Kostas Krikellas | 03 April 2024, 13:21:41 UTC | [TEST] Add rolling upgrade test for downsampling (#107036) Simple test, sets up downsampling to run in the old cluster, then waits for it to complete and verifies that downsampled indexes can get queried in the mixed and upgraded clusters. | 03 April 2024, 13:21:41 UTC |
| febc23c | Pat Whelan | 03 April 2024, 13:15:19 UTC | [Transform] Clarify afterFinishOrFailure javadoc (#107048) Close #100515 | 03 April 2024, 13:15:19 UTC |
| bb520eb | Przemysław Witek | 03 April 2024, 13:07:48 UTC | [Transform] Implement robustness test that constantly creates/starts/stops/deletes a continuous transform (#106313) | 03 April 2024, 13:07:48 UTC |
| 1c923a6 | Pat Whelan | 03 April 2024, 11:46:12 UTC | [ML] Convert DatafeedContext to a Record (#107009) Refactor DatafeedContext from a POJO with a Builder to a Record. This is mostly to reduce code footprint and improve readability in the `DatafeedContextProvider` class. The lambda closures are now responsible for passing the fields to the DatafeedContext, rather than the Builder. Close #107008 | 03 April 2024, 11:46:12 UTC |
| c74490c | Nik Everett | 03 April 2024, 11:42:40 UTC | ESQL: Enable VALUES agg for datetime (#107016) When I wrote the `VALUES` agg I didn't plug it in for `datetime` fields. Ooops. We just have to plug it in. | 03 April 2024, 11:42:40 UTC |
| aa03bc8 | Bogdan Pintea | 03 April 2024, 11:30:09 UTC | ESQL: Swap `FROM` `METADATA` and `OPTIONS` directives (#107012) This swaps the two clauses, foreseeing a more frequent use of METADATA one, over OPTIONS. | 03 April 2024, 11:30:09 UTC |
| a23e29c | Yang Wang | 03 April 2024, 09:34:58 UTC | [Test] Explicitly specify number of shards to 1 (#107032) The test assumes number of shards defaults to 1 which may not true in certain deployment types. This PR makes it explicity configure it to 1. Relates: #106707 | 03 April 2024, 09:34:58 UTC |
| 331d783 | István Zoltán Szabó | 03 April 2024, 09:30:14 UTC | [DOCS] Fixes search request in semantic search tutorial. (#107034) | 03 April 2024, 09:30:14 UTC |
| 0c493e4 | Luca Cavanna | 03 April 2024, 09:21:06 UTC | Make requestSize final in SearchExecutionContext (#107026) | 03 April 2024, 09:21:06 UTC |
| 573c032 | Liam Thompson | 03 April 2024, 08:54:49 UTC | [Docs] Fix CCS matrix for 8.13 (#107028) | 03 April 2024, 08:54:49 UTC |
| 99138c1 | Mary Gouseti | 03 April 2024, 08:14:47 UTC | [Failure store] Refactoring data stream lifecycle to facilitate managing failure store. (#106981) In this PR we refactor data stream and data stream lifecycle related code. The goal of this small refactoring is to make it easier to extend DSL to manage failure store. - The retrieval of backing indices past the retention period. - The rollover execution during the DSL run to handle more than one write index. | 03 April 2024, 08:14:47 UTC |
| 80bba1a | Alexander Spies | 03 April 2024, 07:58:52 UTC | ESQL: Use SearchExecutionContext in SearchStats (#106999) This is a lighter dependency than requiring SearchContext. | 03 April 2024, 07:58:52 UTC |
| 431f73f | Kostas Krikellas | 03 April 2024, 06:02:23 UTC | [TEST] Wait for downsampling op to complete on master side (#106918) * unmute test, remove failing assert * Wait for downsampling op to complete on master side * read asserted value properly * add invalid config status | 03 April 2024, 06:02:23 UTC |
| f930544 | Nhat Nguyen | 03 April 2024, 02:35:48 UTC | Introduce ordinal bytesref block (#106852) This PR introduces ordinal-based blocks for BytesRef. These blocks consist of a pair: an IntBlock for ordinals and a BytesRefVector for the dictionary. Compared to the regular BytesRefBlock, these blocks are slower due to indirect access and consume more memory because of the additional ordinals block. However, they offer significant speed improvements and reduced memory usage when byte values are frequently repeated. There are several use cases where these blocks can be beneficial. | 03 April 2024, 02:35:48 UTC |
| e6ccaf4 | Nik Everett | 03 April 2024, 00:42:14 UTC | ESQL: Split the tests for META FUNCTIONS (#106954) This splits the tests for META FUNCTIONS into a few tests that fit better on the screen. There's an interesting bug in the CSV parser where if you end the last line with an empty string it'll consider it `null` which we think of as `null`. I bumped into that and it's caused me some trouble. I had to work around it here. Thus the assertion changes. Basically `[]` will force us to parse a list. So we can use `[""]` - a list of just the empty string - to force the testing framework to spit out an empty string. We only need this because the empty string is the last item in the list.... | 03 April 2024, 00:42:14 UTC |
| e39fd58 | Ryan Ernst | 02 April 2024, 20:40:33 UTC | Fix jvm ergonomics tests (#106969) closes #106554 | 02 April 2024, 20:40:33 UTC |
| 1e253a0 | Volodymyr Krasnikov | 02 April 2024, 20:39:01 UTC | Propagate ?master_timeout query parameter from CCR apis to downstreams (#105168) * Propagate ?master_timeout query parameter from crr put follow api * Update docs/changelog/105168.yaml * fix sily mistake * spotless * fix 2 * Add ?master_timeout query param to apis which use it * Update rest api specs + yaml tests * Add master_timeout (+timeout) to remaining api endpoints * Update 105168.yaml Correct message * Enable randomly ?master_timeout param to ccr IT tests * Add timeout param to RestCcrStatsAction * propagate master_timeout param between put_follow -> resume_follow calls * Propagate master_timeout down to persistent task layer * Add transport version for ccr stats request object change * Add BwC test for CcrStatsAction.Request obj | 02 April 2024, 20:39:01 UTC |
| 412f2f5 | Daniel | 02 April 2024, 19:53:19 UTC | New Lucene spanish plural stemmer (#106952) Add Lucene SpanishPluralStemFilter based on apache/lucene#11284 | 02 April 2024, 19:53:19 UTC |
| 646ac4f | Efe Gürkan YALAMAN | 02 April 2024, 17:20:39 UTC | Update license acknowledgement messages for Search (#106857) License state messages are shown to user when license expires, or license state changed. Updated Search text to include current features. | 02 April 2024, 17:20:39 UTC |
| ee7e254 | Athena Brown | 02 April 2024, 17:14:41 UTC | Remove remaining Version usages in licensing (#101800) This commit is a follow-up to the changes converting the trial license refresh logic to use `TrialLicenseVersion` (#100169), converting the remaining uses of `Version`, which are used to ensure wire compatibility, with `TransportVersion`. Incidentally, this commit also removes code for compatibility with 7.6.0 and earlier, as those versions can no longer be in a cluster with current-version nodes. | 02 April 2024, 17:14:41 UTC |
| 08c7b65 | David Turner | 02 April 2024, 15:14:22 UTC | Remove `ClusterStateObserver` constructor with implicit timeout (#107004) This constructor is only used in tests, and it's trappy to hide the timeout from callers like this. Inlining it to make the timeout explicit everywhere. | 02 April 2024, 15:14:22 UTC |
| 499ebbf | Craig Taverner | 02 April 2024, 14:36:56 UTC | ESQL: Fix issues with ST_* functions and compound geometries (#106992) * Add mores tests to cartesian complex geometry tests * Support multi-shapes within constant literals This was not supported for lucene push-down for any of the relates functions on point indices: ST_INTERSECTS, ST_WITHIN, ST_CONTAINS. In addition ST_CONTAINS had specific bugs around handling multi-component shapes within literals/constants. * Update docs/changelog/106992.yaml * Delete docs/changelog/106992.yaml * Control result order for multi-cluster tests | 02 April 2024, 14:36:56 UTC |
| e8a3c9b | Mark Vieira | 02 April 2024, 14:12:42 UTC | AwaitsFix #106968 | 02 April 2024, 14:12:42 UTC |
| 225bb2c | Nikolaj Volgushev | 02 April 2024, 13:58:13 UTC | Inject reserved role name checker into role mgmt actions (#106759) This PR makes reserved role name validation injectable for the create and delete roles actions. | 02 April 2024, 13:58:13 UTC |
| 959ef23 | David Turner | 02 April 2024, 13:51:56 UTC | Accumulate results directly in TransportGetSnapshotsAction (#106983) There's no need to return a `SnapshotsInRepo` object all the way along the chain of listeners, we can just collect the results directly and signal completion with a `Void` response. | 02 April 2024, 13:51:56 UTC |
| d8a9046 | Nik Everett | 02 April 2024, 13:45:35 UTC | ESQL: Fix VALUES test for `text` (#106955) This fixes an error in the test for `text` fields running the `VALUES` agg. We have to sort the output or we'll get it in a random order. Closes #106939 Closes #106894 | 02 April 2024, 13:45:35 UTC |
| 3c10d47 | elasticsearchmachine | 02 April 2024, 13:38:50 UTC | Prune changelogs after 8.13.1 release | 02 April 2024, 13:38:50 UTC |
| 35b7212 | elasticsearchmachine | 02 April 2024, 13:37:40 UTC | Bump versions after 8.13.1 release | 02 April 2024, 13:37:41 UTC |
| c9ab203 | Mark Vieira | 02 April 2024, 13:22:54 UTC | Remove AwaitsFix | 02 April 2024, 13:22:54 UTC |
| f2e594f | Armin Braun | 02 April 2024, 13:18:54 UTC | Remove unused XContent parsing logic (#106950) Another round of mostly automatically cleaning up unused xcontent parser methods and parser instances that become unused as a result. | 02 April 2024, 13:18:54 UTC |
| ee6ff33 | Nik Everett | 02 April 2024, 12:46:41 UTC | ESQL: Remove unused test class (#106996) We don't use it any more. | 02 April 2024, 12:46:41 UTC |
| 480ee1c | Luca Cavanna | 02 April 2024, 12:38:59 UTC | Convert FetchFieldsContext to a record (#106991) Simple refactoring to save a few lines of code. | 02 April 2024, 12:38:59 UTC |
| e56dcee | Tim Grein | 02 April 2024, 11:34:46 UTC | [Inference API] Add completion task type docs (#106876) | 02 April 2024, 11:34:46 UTC |
| 7f17eff | Niels Bauman | 02 April 2024, 11:03:35 UTC | Implement rollover for failure stores (#106715) Allows rolling over failure stores through the existing RolloverRequest by specifying `IndicesOptions.FailureStoreOptions` on that request. | 02 April 2024, 11:03:35 UTC |
| 2380492 | Craig Taverner | 02 April 2024, 08:31:00 UTC | ESQL: Support ST_CONTAINS and ST_WITHIN (#106503) * WIP Started adding ST_CONTAINS * Add generated evaluators * Reduced warnings and use correct evaluators * Refactored tests to remove duplicate code, and fixed Contains/multi-components * Gradle build disallows using getDeclaredField * Fixed cases where rectangles cross the dateline * Fixed meta function tests * Added ST_WITHIN to support inverting ST_CONTAINS If the ST_CONTAINS is called with the constant on the left, we either have to create a lot more Evaluators to cover that case, or we have to invert it to ST_WITHIN. This inversion was a much easier option. * Simplify inversion logic * Add comment on choice of surrogate approach * Add unit tests and missing fold() function * Simple code cleanup * Add integration tests for literals * Add more integration tests based on actual data * Generated documentation files * Add documentation * Fixed failing function count test * Add tests that push-to-source works for ST_CONTAINS and ST_WITHIN * Test more combinations of WITH/CONTAINS and literal on right and left This also verifies that the re-writing of CONTAINS to WITHIN or vice versa occurs when the literal is on the left. * test that physical planning also handles doc-values from STATS * Added more tests for WITHIN/CONTAINS together with CENTROID This should test the doc-values for points. * Add cartesian_point tests * Add cartesian_shape tests * Disable Lucene-push-down for CARTESIAN data This is a limitation in Lucene, which we could address as a performance optimization in a future PR, but since it probably requires Lucene changes, it cannot be done in this work. * Fix doc links * Added test data and tests for cartesian multi-polygons Testing INTERSECTS, CONTAINS and WITHIN with multi-polydon fields * Use required features for spatial points, shapes and centroid * 8.13.0 is not yet historical version This needs to be reverted as soon as 8.13.0 is released * Added st_intersects and st_contains_within 'features' * Code review updates * Re-enable lucene push-down * Added more required_features * Fix point contains non-point * Fix point contains point * Re-enable lucene push-down in tests too Forgot to change the physical planner unit tests after re-enabling lucene push-down * Generate automatic docs * Use generated examples docs * Generated examples use '-result' prefix (singular) * Mark spatial functions as preview/experimental | 02 April 2024, 08:31:00 UTC |
| 20fc2d2 | Tim Vernum | 02 April 2024, 08:27:33 UTC | Test modifying and removing data in ProfileIT (#106582) User Profiles can be used to store application data against a user (e.g. user preferences). This commit extends the integration tests for profile date storage to include explict tests for partial updates and clearing existing data | 02 April 2024, 08:27:33 UTC |
| 7ee63df | David Turner | 02 April 2024, 08:18:59 UTC | Expand docs on reset desired balance API (#106921) Explains its purpose and surrounding context a little more, including a note that this should never be necessary (i.e. if you find you need it, that's a bug). | 02 April 2024, 08:18:59 UTC |
| 519580e | David Turner | 02 April 2024, 07:42:07 UTC | AwaitsFix for #106957 | 02 April 2024, 07:42:19 UTC |
| 0eca03c | David Turner | 02 April 2024, 06:52:25 UTC | Remove unused `SnapshotsInRepo#remaining` (#106674) We only discard snapshots using the `?size` parameter when constructing the final response, so we can count the `remaining` snapshots in a local variable rather than tracking an unnecessary `0` on every `SnapshotsInRepo`. Indeed by inlining `sortSnapshots` into the routine that constructs the final response we can avoid even having to build the a final `SnapshotsInRepo` here. | 02 April 2024, 06:52:25 UTC |
| 89cfb85 | Albert Zaharovits | 02 April 2024, 06:36:50 UTC | [Test] Fix AsyncSearchSecurityIT testStatusWithUsersWhileSearchIsRunning (#106912) The error_query is only available in snapshot builds. All test failures have the release-tests tag. Closes #106871 | 02 April 2024, 06:36:50 UTC |
| df0fd30 | Albert Zaharovits | 02 April 2024, 06:35:02 UTC | [Doc] Privileges required to retrieve the status of async searches Document that users can retrieve the status of the async searches they submitted without any extra privileges. | 02 April 2024, 06:35:02 UTC |
| 4ddb640 | Armin Braun | 02 April 2024, 00:41:22 UTC | Remove o.e.painless.toxcontent.UserTreeToXContent (#106935) This was never used -> remove it and its test. | 02 April 2024, 00:41:22 UTC |
| 393e644 | Armin Braun | 01 April 2024, 22:44:44 UTC | Refactor more ActionListener.wrap to delegateFailureAndWrap (#106948) Refactoring a couple more of these and inlining some listeners while at it to save code/memory and get cleaner guarantees around note leaking things. Note: In two cases inlining makes it obvious that we can remove spurious AtomicReference usage. | 01 April 2024, 22:44:44 UTC |
| 22ca5ec | Lee Hinman | 01 April 2024, 22:19:59 UTC | Fix warning typo for test failure (#106971) This had the wrong index patterns, which led to test failures Resolves #106965 | 01 April 2024, 22:19:59 UTC |
| 6011763 | Lee Hinman | 01 April 2024, 21:52:27 UTC | Call out `monitor` privilege for index and component templates (#106970) These can be retrieved using the `monitor` cluster privilege. This just adds that note to the docs. | 01 April 2024, 21:52:27 UTC |
| 5cc4a56 | Mark Vieira | 01 April 2024, 21:09:52 UTC | AwaitsFix #106939 | 01 April 2024, 21:09:52 UTC |
| 9077c5d | Mark Vieira | 01 April 2024, 21:04:56 UTC | AwaitsFix #106968 | 01 April 2024, 21:04:56 UTC |
| f9d96ae | Mark Vieira | 01 April 2024, 20:44:21 UTC | AwaitsFix #106900 | 01 April 2024, 20:44:31 UTC |
| 8fdd1aa | Joe Gallo | 01 April 2024, 20:34:15 UTC | Refactor geoip database properties (#106960) | 01 April 2024, 20:34:15 UTC |
| 3eeffeb | Mark Vieira | 01 April 2024, 20:33:22 UTC | AwaitsFix #106964 | 01 April 2024, 20:33:22 UTC |
| a3794e7 | shainaraskas | 01 April 2024, 20:12:48 UTC | [DOCS] Remove orphaned cluster issues troubleshooing doc (#106959) | 01 April 2024, 20:12:48 UTC |
| 3810a91 | Ryan Ernst | 01 April 2024, 19:49:29 UTC | Block readiness on file settings being applied (#106437) The readiness service is supposed to wait on the cluster being formed as well as file settings being applied. Yet file settings application is only checked on the master node. Since master election itself does not guarantee file settings have been applied, non-master nodes may mark themselves as ready even though they should not yet accept requests. This commit reworks how the readiness service looks ofr file settings being applied. Previously it only worked on the master node where it got a callback directly from the file settings service. With this change we now only look at cluster state. Note that this means the readiness service and file based settings are tightly coupled. In practice this was always the case as they are both meant to be used in cloud environments, but the coupling is a bit tighter now since the readiness service will never report ready unless file based settings exist. | 01 April 2024, 19:49:29 UTC |
