Lock manager: Fix setup bug #230519
Merged
Lock manager: Fix setup bug #230519
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Contributor
|
Pinging @elastic/obs-knowledge-team (Team:obs-knowledge) |
…sh1293/kibana into flash1293/fix-lock-manager-setup-bug
arturoliduena
approved these changes
Aug 5, 2025
SrdjanLL
approved these changes
Aug 5, 2025
Contributor
SrdjanLL
left a comment
There was a problem hiding this comment.
LGTM, just left a thought on how this new behaviour may impact downstream.
| let runSetupIndexAssetOnce = once(setupLockManagerIndex); | ||
| export function runSetupIndexAssetEveryTime() { | ||
| runSetupIndexAssetOnce = setupLockManagerIndex; | ||
| export function rerunSetupIndexAsset() { |
Contributor
There was a problem hiding this comment.
[Observation] Do you think there are any risks of exposing this functionality to downstream dependencies of the lock manager? Guess it was here before, which was going to cache the failures, but now that it's passing through, is this exposing the "reset" functionality downstream?
Contributor
Author
There was a problem hiding this comment.
It shouldn't change anything, it's just the new version of runSetIndexAssetEveryTime, but with a new name.
Contributor
|
Starting backport for target branches: 9.1 |
Contributor
💚 Build Succeeded
Metrics [docs]
History
|
kibanamachine
pushed a commit
to kibanamachine/kibana
that referenced
this pull request
Aug 5, 2025
Fixes elastic#230499 The lock manager runs `setupLockManagerIndex` via lodash `once`, so it's not happening on every call. However, if the first call to `setupLockManagerIndex` errors out (e.g. because Elasticsearch isn't ready yet), then every subsequent call will return the cached rejected promise and fail as well, rendering all lock managers in that node instance broken (since `once` keeps its state on module scope) This leads to issues like this (timeout exception thrown from streams, but the call stack originates from the slo plugin setup routine since it's the cached rejected promise): ``` [2025-08-01T18:36:25.080+00:00][ERROR][plugins.streams] TimeoutError: Request timed out at KibanaTransport._request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:564:50) at processTicksAndRejections (node:internal/process/task_queues:105:5) at runNextTicks (node:internal/process/task_queues:69:3) at listOnTimeout (node:internal/timers:549:9) at processTimers (node:internal/timers:523:7) at /usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:631:32 at KibanaTransport.request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:627:20) at KibanaTransport.request (/usr/share/kibana/node_modules/@kbn/core-elasticsearch-client-server-internal/src/create_transport.js:60:16) at Cluster.putComponentTemplate (/usr/share/kibana/node_modules/@elastic/elasticsearch/lib/api/api/cluster.js:600:16) at ensureTemplatesAndIndexCreated (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:56:3) at setupLockManagerIndex (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:110:3) at LockManager.acquire (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:53:5) at withLock (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:242:20) at /usr/share/kibana/node_modules/@kbn/slo-plugin/server/plugin.js:176:7 ``` This PR fixes the problem by not using once but instead keeping the state manually only if the promise succeeds, passing errors through. (cherry picked from commit b4f8488)
Contributor
💚 All backports created successfully
Note: Successful backport PRs will be merged automatically after passing CI. Questions ?Please refer to the Backport tool documentation |
kibanamachine
added a commit
that referenced
this pull request
Aug 5, 2025
# Backport This will backport the following commits from `main` to `9.1`: - [Lock manager: Fix setup bug (#230519)](#230519) <!--- Backport version: 9.6.6 --> ### Questions ? Please refer to the [Backport tool documentation](https://github.com/sorenlouv/backport) <!--BACKPORT [{"author":{"name":"Joe Reuter","email":"[email protected]"},"sourceCommit":{"committedDate":"2025-08-05T11:41:28Z","message":"Lock manager: Fix setup bug (#230519)\n\nFixes https://github.com/elastic/kibana/issues/230499\n\nThe lock manager runs `setupLockManagerIndex` via lodash `once`, so it's\nnot happening on every call. However, if the first call to\n`setupLockManagerIndex` errors out (e.g. because Elasticsearch isn't\nready yet), then every subsequent call will return the cached rejected\npromise and fail as well, rendering all lock managers in that node\ninstance broken (since `once` keeps its state on module scope)\n\nThis leads to issues like this (timeout exception thrown from streams,\nbut the call stack originates from the slo plugin setup routine since\nit's the cached rejected promise):\n```\n[2025-08-01T18:36:25.080+00:00][ERROR][plugins.streams] TimeoutError: Request timed out\n at KibanaTransport._request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:564:50)\n at processTicksAndRejections (node:internal/process/task_queues:105:5)\n at runNextTicks (node:internal/process/task_queues:69:3)\n at listOnTimeout (node:internal/timers:549:9)\n at processTimers (node:internal/timers:523:7)\n at /usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:631:32\n at KibanaTransport.request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:627:20)\n at KibanaTransport.request (/usr/share/kibana/node_modules/@kbn/core-elasticsearch-client-server-internal/src/create_transport.js:60:16)\n at Cluster.putComponentTemplate (/usr/share/kibana/node_modules/@elastic/elasticsearch/lib/api/api/cluster.js:600:16)\n at ensureTemplatesAndIndexCreated (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:56:3)\n at setupLockManagerIndex (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:110:3)\n at LockManager.acquire (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:53:5)\n at withLock (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:242:20)\n at /usr/share/kibana/node_modules/@kbn/slo-plugin/server/plugin.js:176:7\n```\n\n\nThis PR fixes the problem by not using once but instead keeping the\nstate manually only if the promise succeeds, passing errors through.","sha":"b4f8488f6c8d3758797e2b0efde3c67510b9707d","branchLabelMapping":{"^v9.2.0$":"main","^v(\\d+).(\\d+).\\d+$":"$1.$2"}},"sourcePullRequest":{"labels":["release_note:fix","Team:obs-knowledge","backport:version","v9.2.0","v9.1.1"],"title":"Lock manager: Fix setup bug","number":230519,"url":"https://github.com/elastic/kibana/pull/230519","mergeCommit":{"message":"Lock manager: Fix setup bug (#230519)\n\nFixes https://github.com/elastic/kibana/issues/230499\n\nThe lock manager runs `setupLockManagerIndex` via lodash `once`, so it's\nnot happening on every call. However, if the first call to\n`setupLockManagerIndex` errors out (e.g. because Elasticsearch isn't\nready yet), then every subsequent call will return the cached rejected\npromise and fail as well, rendering all lock managers in that node\ninstance broken (since `once` keeps its state on module scope)\n\nThis leads to issues like this (timeout exception thrown from streams,\nbut the call stack originates from the slo plugin setup routine since\nit's the cached rejected promise):\n```\n[2025-08-01T18:36:25.080+00:00][ERROR][plugins.streams] TimeoutError: Request timed out\n at KibanaTransport._request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:564:50)\n at processTicksAndRejections (node:internal/process/task_queues:105:5)\n at runNextTicks (node:internal/process/task_queues:69:3)\n at listOnTimeout (node:internal/timers:549:9)\n at processTimers (node:internal/timers:523:7)\n at /usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:631:32\n at KibanaTransport.request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:627:20)\n at KibanaTransport.request (/usr/share/kibana/node_modules/@kbn/core-elasticsearch-client-server-internal/src/create_transport.js:60:16)\n at Cluster.putComponentTemplate (/usr/share/kibana/node_modules/@elastic/elasticsearch/lib/api/api/cluster.js:600:16)\n at ensureTemplatesAndIndexCreated (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:56:3)\n at setupLockManagerIndex (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:110:3)\n at LockManager.acquire (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:53:5)\n at withLock (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:242:20)\n at /usr/share/kibana/node_modules/@kbn/slo-plugin/server/plugin.js:176:7\n```\n\n\nThis PR fixes the problem by not using once but instead keeping the\nstate manually only if the promise succeeds, passing errors through.","sha":"b4f8488f6c8d3758797e2b0efde3c67510b9707d"}},"sourceBranch":"main","suggestedTargetBranches":["9.1"],"targetPullRequestStates":[{"branch":"main","label":"v9.2.0","branchLabelMappingKey":"^v9.2.0$","isSourceBranch":true,"state":"MERGED","url":"https://github.com/elastic/kibana/pull/230519","number":230519,"mergeCommit":{"message":"Lock manager: Fix setup bug (#230519)\n\nFixes https://github.com/elastic/kibana/issues/230499\n\nThe lock manager runs `setupLockManagerIndex` via lodash `once`, so it's\nnot happening on every call. However, if the first call to\n`setupLockManagerIndex` errors out (e.g. because Elasticsearch isn't\nready yet), then every subsequent call will return the cached rejected\npromise and fail as well, rendering all lock managers in that node\ninstance broken (since `once` keeps its state on module scope)\n\nThis leads to issues like this (timeout exception thrown from streams,\nbut the call stack originates from the slo plugin setup routine since\nit's the cached rejected promise):\n```\n[2025-08-01T18:36:25.080+00:00][ERROR][plugins.streams] TimeoutError: Request timed out\n at KibanaTransport._request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:564:50)\n at processTicksAndRejections (node:internal/process/task_queues:105:5)\n at runNextTicks (node:internal/process/task_queues:69:3)\n at listOnTimeout (node:internal/timers:549:9)\n at processTimers (node:internal/timers:523:7)\n at /usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:631:32\n at KibanaTransport.request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:627:20)\n at KibanaTransport.request (/usr/share/kibana/node_modules/@kbn/core-elasticsearch-client-server-internal/src/create_transport.js:60:16)\n at Cluster.putComponentTemplate (/usr/share/kibana/node_modules/@elastic/elasticsearch/lib/api/api/cluster.js:600:16)\n at ensureTemplatesAndIndexCreated (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:56:3)\n at setupLockManagerIndex (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:110:3)\n at LockManager.acquire (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:53:5)\n at withLock (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:242:20)\n at /usr/share/kibana/node_modules/@kbn/slo-plugin/server/plugin.js:176:7\n```\n\n\nThis PR fixes the problem by not using once but instead keeping the\nstate manually only if the promise succeeds, passing errors through.","sha":"b4f8488f6c8d3758797e2b0efde3c67510b9707d"}},{"branch":"9.1","label":"v9.1.1","branchLabelMappingKey":"^v(\\d+).(\\d+).\\d+$","isSourceBranch":false,"state":"NOT_CREATED"}]}] BACKPORT--> Co-authored-by: Joe Reuter <[email protected]>
delanni
pushed a commit
to delanni/kibana
that referenced
this pull request
Aug 5, 2025
Fixes elastic#230499 The lock manager runs `setupLockManagerIndex` via lodash `once`, so it's not happening on every call. However, if the first call to `setupLockManagerIndex` errors out (e.g. because Elasticsearch isn't ready yet), then every subsequent call will return the cached rejected promise and fail as well, rendering all lock managers in that node instance broken (since `once` keeps its state on module scope) This leads to issues like this (timeout exception thrown from streams, but the call stack originates from the slo plugin setup routine since it's the cached rejected promise): ``` [2025-08-01T18:36:25.080+00:00][ERROR][plugins.streams] TimeoutError: Request timed out at KibanaTransport._request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:564:50) at processTicksAndRejections (node:internal/process/task_queues:105:5) at runNextTicks (node:internal/process/task_queues:69:3) at listOnTimeout (node:internal/timers:549:9) at processTimers (node:internal/timers:523:7) at /usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:631:32 at KibanaTransport.request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:627:20) at KibanaTransport.request (/usr/share/kibana/node_modules/@kbn/core-elasticsearch-client-server-internal/src/create_transport.js:60:16) at Cluster.putComponentTemplate (/usr/share/kibana/node_modules/@elastic/elasticsearch/lib/api/api/cluster.js:600:16) at ensureTemplatesAndIndexCreated (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:56:3) at setupLockManagerIndex (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:110:3) at LockManager.acquire (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:53:5) at withLock (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:242:20) at /usr/share/kibana/node_modules/@kbn/slo-plugin/server/plugin.js:176:7 ``` This PR fixes the problem by not using once but instead keeping the state manually only if the promise succeeds, passing errors through.
Contributor
Author
💚 All backports created successfully
Note: Successful backport PRs will be merged automatically after passing CI. Questions ?Please refer to the Backport tool documentation |
flash1293
added a commit
to flash1293/kibana
that referenced
this pull request
Aug 6, 2025
Fixes elastic#230499 The lock manager runs `setupLockManagerIndex` via lodash `once`, so it's not happening on every call. However, if the first call to `setupLockManagerIndex` errors out (e.g. because Elasticsearch isn't ready yet), then every subsequent call will return the cached rejected promise and fail as well, rendering all lock managers in that node instance broken (since `once` keeps its state on module scope) This leads to issues like this (timeout exception thrown from streams, but the call stack originates from the slo plugin setup routine since it's the cached rejected promise): ``` [2025-08-01T18:36:25.080+00:00][ERROR][plugins.streams] TimeoutError: Request timed out at KibanaTransport._request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:564:50) at processTicksAndRejections (node:internal/process/task_queues:105:5) at runNextTicks (node:internal/process/task_queues:69:3) at listOnTimeout (node:internal/timers:549:9) at processTimers (node:internal/timers:523:7) at /usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:631:32 at KibanaTransport.request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:627:20) at KibanaTransport.request (/usr/share/kibana/node_modules/@kbn/core-elasticsearch-client-server-internal/src/create_transport.js:60:16) at Cluster.putComponentTemplate (/usr/share/kibana/node_modules/@elastic/elasticsearch/lib/api/api/cluster.js:600:16) at ensureTemplatesAndIndexCreated (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:56:3) at setupLockManagerIndex (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:110:3) at LockManager.acquire (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:53:5) at withLock (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:242:20) at /usr/share/kibana/node_modules/@kbn/slo-plugin/server/plugin.js:176:7 ``` This PR fixes the problem by not using once but instead keeping the state manually only if the promise succeeds, passing errors through. (cherry picked from commit b4f8488)
flash1293
added a commit
that referenced
this pull request
Aug 6, 2025
# Backport This will backport the following commits from `main` to `8.19`: - [Lock manager: Fix setup bug (#230519)](#230519) <!--- Backport version: 10.0.1 --> ### Questions ? Please refer to the [Backport tool documentation](https://github.com/sorenlouv/backport) <!--BACKPORT [{"author":{"name":"Joe Reuter","email":"[email protected]"},"sourceCommit":{"committedDate":"2025-08-05T11:41:28Z","message":"Lock manager: Fix setup bug (#230519)\n\nFixes https://github.com/elastic/kibana/issues/230499\n\nThe lock manager runs `setupLockManagerIndex` via lodash `once`, so it's\nnot happening on every call. However, if the first call to\n`setupLockManagerIndex` errors out (e.g. because Elasticsearch isn't\nready yet), then every subsequent call will return the cached rejected\npromise and fail as well, rendering all lock managers in that node\ninstance broken (since `once` keeps its state on module scope)\n\nThis leads to issues like this (timeout exception thrown from streams,\nbut the call stack originates from the slo plugin setup routine since\nit's the cached rejected promise):\n```\n[2025-08-01T18:36:25.080+00:00][ERROR][plugins.streams] TimeoutError: Request timed out\n at KibanaTransport._request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:564:50)\n at processTicksAndRejections (node:internal/process/task_queues:105:5)\n at runNextTicks (node:internal/process/task_queues:69:3)\n at listOnTimeout (node:internal/timers:549:9)\n at processTimers (node:internal/timers:523:7)\n at /usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:631:32\n at KibanaTransport.request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:627:20)\n at KibanaTransport.request (/usr/share/kibana/node_modules/@kbn/core-elasticsearch-client-server-internal/src/create_transport.js:60:16)\n at Cluster.putComponentTemplate (/usr/share/kibana/node_modules/@elastic/elasticsearch/lib/api/api/cluster.js:600:16)\n at ensureTemplatesAndIndexCreated (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:56:3)\n at setupLockManagerIndex (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:110:3)\n at LockManager.acquire (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:53:5)\n at withLock (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:242:20)\n at /usr/share/kibana/node_modules/@kbn/slo-plugin/server/plugin.js:176:7\n```\n\n\nThis PR fixes the problem by not using once but instead keeping the\nstate manually only if the promise succeeds, passing errors through.","sha":"b4f8488f6c8d3758797e2b0efde3c67510b9707d","branchLabelMapping":{"^v9.2.0$":"main","^v(\\d+).(\\d+).\\d+$":"$1.$2"}},"sourcePullRequest":{"labels":["release_note:fix","Team:obs-knowledge","backport:version","v9.2.0","v9.1.1","v8.19.1"],"title":"Lock manager: Fix setup bug","number":230519,"url":"https://github.com/elastic/kibana/pull/230519","mergeCommit":{"message":"Lock manager: Fix setup bug (#230519)\n\nFixes https://github.com/elastic/kibana/issues/230499\n\nThe lock manager runs `setupLockManagerIndex` via lodash `once`, so it's\nnot happening on every call. However, if the first call to\n`setupLockManagerIndex` errors out (e.g. because Elasticsearch isn't\nready yet), then every subsequent call will return the cached rejected\npromise and fail as well, rendering all lock managers in that node\ninstance broken (since `once` keeps its state on module scope)\n\nThis leads to issues like this (timeout exception thrown from streams,\nbut the call stack originates from the slo plugin setup routine since\nit's the cached rejected promise):\n```\n[2025-08-01T18:36:25.080+00:00][ERROR][plugins.streams] TimeoutError: Request timed out\n at KibanaTransport._request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:564:50)\n at processTicksAndRejections (node:internal/process/task_queues:105:5)\n at runNextTicks (node:internal/process/task_queues:69:3)\n at listOnTimeout (node:internal/timers:549:9)\n at processTimers (node:internal/timers:523:7)\n at /usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:631:32\n at KibanaTransport.request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:627:20)\n at KibanaTransport.request (/usr/share/kibana/node_modules/@kbn/core-elasticsearch-client-server-internal/src/create_transport.js:60:16)\n at Cluster.putComponentTemplate (/usr/share/kibana/node_modules/@elastic/elasticsearch/lib/api/api/cluster.js:600:16)\n at ensureTemplatesAndIndexCreated (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:56:3)\n at setupLockManagerIndex (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:110:3)\n at LockManager.acquire (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:53:5)\n at withLock (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:242:20)\n at /usr/share/kibana/node_modules/@kbn/slo-plugin/server/plugin.js:176:7\n```\n\n\nThis PR fixes the problem by not using once but instead keeping the\nstate manually only if the promise succeeds, passing errors through.","sha":"b4f8488f6c8d3758797e2b0efde3c67510b9707d"}},"sourceBranch":"main","suggestedTargetBranches":["8.19"],"targetPullRequestStates":[{"branch":"main","label":"v9.2.0","branchLabelMappingKey":"^v9.2.0$","isSourceBranch":true,"state":"MERGED","url":"https://github.com/elastic/kibana/pull/230519","number":230519,"mergeCommit":{"message":"Lock manager: Fix setup bug (#230519)\n\nFixes https://github.com/elastic/kibana/issues/230499\n\nThe lock manager runs `setupLockManagerIndex` via lodash `once`, so it's\nnot happening on every call. However, if the first call to\n`setupLockManagerIndex` errors out (e.g. because Elasticsearch isn't\nready yet), then every subsequent call will return the cached rejected\npromise and fail as well, rendering all lock managers in that node\ninstance broken (since `once` keeps its state on module scope)\n\nThis leads to issues like this (timeout exception thrown from streams,\nbut the call stack originates from the slo plugin setup routine since\nit's the cached rejected promise):\n```\n[2025-08-01T18:36:25.080+00:00][ERROR][plugins.streams] TimeoutError: Request timed out\n at KibanaTransport._request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:564:50)\n at processTicksAndRejections (node:internal/process/task_queues:105:5)\n at runNextTicks (node:internal/process/task_queues:69:3)\n at listOnTimeout (node:internal/timers:549:9)\n at processTimers (node:internal/timers:523:7)\n at /usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:631:32\n at KibanaTransport.request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:627:20)\n at KibanaTransport.request (/usr/share/kibana/node_modules/@kbn/core-elasticsearch-client-server-internal/src/create_transport.js:60:16)\n at Cluster.putComponentTemplate (/usr/share/kibana/node_modules/@elastic/elasticsearch/lib/api/api/cluster.js:600:16)\n at ensureTemplatesAndIndexCreated (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:56:3)\n at setupLockManagerIndex (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:110:3)\n at LockManager.acquire (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:53:5)\n at withLock (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:242:20)\n at /usr/share/kibana/node_modules/@kbn/slo-plugin/server/plugin.js:176:7\n```\n\n\nThis PR fixes the problem by not using once but instead keeping the\nstate manually only if the promise succeeds, passing errors through.","sha":"b4f8488f6c8d3758797e2b0efde3c67510b9707d"}},{"branch":"9.1","label":"v9.1.1","branchLabelMappingKey":"^v(\\d+).(\\d+).\\d+$","isSourceBranch":false,"url":"https://github.com/elastic/kibana/pull/230549","number":230549,"state":"MERGED","mergeCommit":{"sha":"bca9737ad0d48d378716b23a0d1c84ca866e164b","message":"[9.1] Lock manager: Fix setup bug (#230519) (#230549)\n\n# Backport\n\nThis will backport the following commits from `main` to `9.1`:\n- [Lock manager: Fix setup bug\n(#230519)](https://github.com/elastic/kibana/pull/230519)\n\n\n\n### Questions ?\nPlease refer to the [Backport tool\ndocumentation](https://github.com/sorenlouv/backport)\n\n\n\nCo-authored-by: Joe Reuter <[email protected]>"}},{"branch":"8.19","label":"v8.19.1","branchLabelMappingKey":"^v(\\d+).(\\d+).\\d+$","isSourceBranch":false,"state":"NOT_CREATED"}]}] BACKPORT-->
NicholasPeretti
pushed a commit
to NicholasPeretti/kibana
that referenced
this pull request
Aug 18, 2025
Fixes elastic#230499 The lock manager runs `setupLockManagerIndex` via lodash `once`, so it's not happening on every call. However, if the first call to `setupLockManagerIndex` errors out (e.g. because Elasticsearch isn't ready yet), then every subsequent call will return the cached rejected promise and fail as well, rendering all lock managers in that node instance broken (since `once` keeps its state on module scope) This leads to issues like this (timeout exception thrown from streams, but the call stack originates from the slo plugin setup routine since it's the cached rejected promise): ``` [2025-08-01T18:36:25.080+00:00][ERROR][plugins.streams] TimeoutError: Request timed out at KibanaTransport._request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:564:50) at processTicksAndRejections (node:internal/process/task_queues:105:5) at runNextTicks (node:internal/process/task_queues:69:3) at listOnTimeout (node:internal/timers:549:9) at processTimers (node:internal/timers:523:7) at /usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:631:32 at KibanaTransport.request (/usr/share/kibana/node_modules/@elastic/transport/lib/Transport.js:627:20) at KibanaTransport.request (/usr/share/kibana/node_modules/@kbn/core-elasticsearch-client-server-internal/src/create_transport.js:60:16) at Cluster.putComponentTemplate (/usr/share/kibana/node_modules/@elastic/elasticsearch/lib/api/api/cluster.js:600:16) at ensureTemplatesAndIndexCreated (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:56:3) at setupLockManagerIndex (/usr/share/kibana/node_modules/@kbn/lock-manager/src/setup_lock_manager_index.js:110:3) at LockManager.acquire (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:53:5) at withLock (/usr/share/kibana/node_modules/@kbn/lock-manager/src/lock_manager_client.js:242:20) at /usr/share/kibana/node_modules/@kbn/slo-plugin/server/plugin.js:176:7 ``` This PR fixes the problem by not using once but instead keeping the state manually only if the promise succeeds, passing errors through.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Fixes #230499
The lock manager runs
setupLockManagerIndexvia lodashonce, so it's not happening on every call. However, if the first call tosetupLockManagerIndexerrors out (e.g. because Elasticsearch isn't ready yet), then every subsequent call will return the cached rejected promise and fail as well, rendering all lock managers in that node instance broken (sinceoncekeeps its state on module scope)This leads to issues like this (timeout exception thrown from streams, but the call stack originates from the slo plugin setup routine since it's the cached rejected promise):
This PR fixes the problem by not using once but instead keeping the state manually only if the promise succeeds, passing errors through.