[Embeddable] Avoid rerendering loop due to search context reload#203150
Merged
dej611 merged 17 commits intoelastic:mainfrom Dec 11, 2024
Merged
[Embeddable] Avoid rerendering loop due to search context reload#203150dej611 merged 17 commits intoelastic:mainfrom
dej611 merged 17 commits intoelastic:mainfrom
Conversation
botelastic
bot
added
the
ci:project-deploy-observability
Contributor
🤖 GitHub commentsExpand to view the GitHub comments
Just comment with:
|
dej611
added
release_note:fix
Team:Presentation
Contributor
|
Pinging @elastic/kibana-presentation (Team:Presentation) |
botelastic
bot
added
the
ci:project-deploy-observability
Contributor
🤖 GitHub commentsExpand to view the GitHub comments
Just comment with:
|
1 task
nreese
approved these changes
Dec 5, 2024
Contributor
nreese
left a comment
nreese
left a comment
There was a problem hiding this comment.
kibana-presentation changes LGTM. useSearchApi is the right places for these fixes
Contributor
Flaky Test Runner Stats🟠 Some tests failed. - kibana-flaky-test-suite-runner#7555[❌] test/functional/apps/discover/group1/config.ts: 0/50 tests passed. |
Heenawter
reviewed
Dec 5, 2024
packages/presentation/presentation_publishing/interfaces/fetch/publishes_unified_search.ts
Outdated
Show resolved
Hide resolved
Contributor
Flaky Test Runner Stats🟠 Some tests failed. - kibana-flaky-test-suite-runner#7557[❌] test/functional/apps/discover/group1/config.ts: 0/50 tests passed. |
botelastic
bot
added
the
Team:actionable-obs
Contributor
|
Pinging @elastic/obs-ux-management-team (Team:obs-ux-management) |
Contributor
Flaky Test Runner Stats🎉 All tests passed! - kibana-flaky-test-suite-runner#7561[✅] test/functional/apps/discover/group1/config.ts: 50/50 tests passed. |
Contributor
|
Hey @ThomThomson
In fact, 1 - The date range changes Shouldn't the first time a searchSessionId is assigned make the charts to fetch data?
True. The page doesn't need to restore the session. However, If I'm not mistaken, the If my memory serves me right, we basically tried to replicate what Dashboards UI does on the auto-refresh. Happy to find alternatives cc: @markov00 |
maryam-saeidi
approved these changes
Dec 9, 2024
Member
maryam-saeidi
left a comment
maryam-saeidi
left a comment
There was a problem hiding this comment.
Worked as expected locally, thanks for the fix ❤️
Member
There was a problem hiding this comment.
Tested following scenarios, LGTM!
| Error in filter query | Error in equation |
|---|---|
 |
 |
Member
There was a problem hiding this comment.
Did you find any example for showing this message, or did you just keep the previous implementation?
Contributor
Author
There was a problem hiding this comment.
As for now I've kept the same implementation. I did not have all the legacy knowledge of why the 2 errors check, so I've just fixed what I've seen not working for now.
crespocarlos
reviewed
Dec 9, 2024
Contributor
crespocarlos
left a comment
crespocarlos
left a comment
There was a problem hiding this comment.
Could focus the changes on removing the searchSessionId only from embeddable usages?
I'll open a ticket on to refactor what is now called use_search_session.ts, but we need the some of the logic that depends on the change of the searchSessionId to control and force new requests
Contributor
There was a problem hiding this comment.
Please keep this file and remove the usages of searchSessionId on Lens charts
Contributor
There was a problem hiding this comment.
We'll refactor this context to be timestamp-based instead of relying on searchSessionId, but removing this now will have consequences beyond the embedabbles. This context is useful to control and force requests
Contributor
There was a problem hiding this comment.
Please leave this context here, it is used to control and force new requests. As said above it will be changed to not rely on searchSessionId.
...lugins/observability_solution/infra/public/components/asset_details/hooks/use_date_picker.ts
Outdated
Show resolved
Hide resolved
markov00
force-pushed
the
fix/202266-2
branch
from
c1563b9 to
219ccb3
Compare
crespocarlos
approved these changes
Dec 10, 2024
Contributor
crespocarlos
left a comment
crespocarlos
left a comment
There was a problem hiding this comment.
LGTM. Thanks for the changes and fixing the problem.
Contributor
|
I think this might solve the problem with the test: dej611#23 |
Contributor
💚 Build Succeeded
Metrics [docs]Async chunks
Page load bundle
History
|
Contributor
Flaky Test Runner Stats🎉 All tests passed! - kibana-flaky-test-suite-runner#7575[✅] test/functional/apps/discover/group1/config.ts: 50/50 tests passed. |
markov00
approved these changes
Dec 11, 2024
Contributor
|
Starting backport for target branches: 8.x |
kibanamachine
pushed a commit
to kibanamachine/kibana
that referenced
this pull request
Dec 11, 2024
…stic#203150) ## Summary Fixes elastic#202266 This PR fixes the underline rerendering issue at the `useSearchApi` hook level, so any embeddable component who uses this hook would benefit from the fix. Ideally the props passed to the Lens component should be memoized, but this assumption would break many integrations as the previous embeddable component did take care of filtering duplicates. To test this: * Go to `Observability > Alerts > Manage rules` and `Add Rule`, pick the `Custom threshold` option and verify the infinite reload does not happen Once fixed this, another problem bubbled up with the brushing use case: when brushing a chart the chart was always one time range step behind. The other bug was due to the `fetch$(api)` function propagating a stale `data` search context who the Lens embeddable was relying on. To solve this other problem the following changes have been applied: * read the `searchSessionId` from the `api` directly (used for `autoRefresh`) * make sure to test the `Refresh` feature with both relative and absolute time ranges * read the `search context` from the `parentApi` directly if implements the `unifiedSearch` API * to test this, brush and check that the final time range matches the correct time range **Note**: the fundamental issue for the latter is `fetch$` not emitting the up-to-date `data` when the parentApi search context updates. The retrieved `data` is stale and one step behind, so it is not reliable. cc @elastic/kibana-presentation As @ThomThomson noticed in his test failure investigation another issue was found in this PR due to the wrong handling of the searchSessionId within the Observability page (for more details see [his analysis](elastic#203150 (comment))). @markov00 and @crespocarlos helped risolve this problem with some additional changes on the Observability side of things: this will lead to some extra searchSessionId to be created, which will be eventually solved by Observability team [shortly moving away from the `searchSessionId` mechanism](elastic#203412) ### Checklist - [x] [Unit or functional tests](https://www.elastic.co/guide/en/kibana/master/development-tests.html) were updated or added to match the most common scenarios --------- Co-authored-by: Marco Vettorello <marco.vettorello@elastic.co> Co-authored-by: Carlos Crespo <crespocarlos@users.noreply.github.com> (cherry picked from commit d4194ba)
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
Dec 11, 2024
#203150) (#203818) # Backport This will backport the following commits from `main` to `8.x`: - [[Embeddable] Avoid rerendering loop due to search context reload (#203150)](#203150) <!--- Backport version: 9.4.3 --> ### Questions ? Please refer to the [Backport tool documentation](https://github.com/sqren/backport) <!--BACKPORT [{"author":{"name":"Marco Liberati","email":"dej611@users.noreply.github.com"},"sourceCommit":{"committedDate":"2024-12-11T15:03:36Z","message":"[Embeddable] Avoid rerendering loop due to search context reload (#203150)\n\n## Summary\r\n\r\nFixes #202266\r\n\r\nThis PR fixes the underline rerendering issue at the `useSearchApi` hook\r\nlevel, so any embeddable component who uses this hook would benefit from\r\nthe fix.\r\n\r\nIdeally the props passed to the Lens component should be memoized, but\r\nthis assumption would break many integrations as the previous embeddable\r\ncomponent did take care of filtering duplicates.\r\nTo test this:\r\n* Go to `Observability > Alerts > Manage rules` and `Add Rule`, pick the\r\n`Custom threshold` option and verify the infinite reload does not happen\r\n\r\nOnce fixed this, another problem bubbled up with the brushing use case:\r\nwhen brushing a chart the chart was always one time range step behind.\r\nThe other bug was due to the `fetch$(api)` function propagating a stale\r\n`data` search context who the Lens embeddable was relying on.\r\nTo solve this other problem the following changes have been applied:\r\n* read the `searchSessionId` from the `api` directly (used for\r\n`autoRefresh`)\r\n* make sure to test the `Refresh` feature with both relative and\r\nabsolute time ranges\r\n* read the `search context` from the `parentApi` directly if implements\r\nthe `unifiedSearch` API\r\n* to test this, brush and check that the final time range matches the\r\ncorrect time range\r\n\r\n**Note**: the fundamental issue for the latter is `fetch# Backport This will backport the following commits from `main` to `8.x`: - [[Embeddable] Avoid rerendering loop due to search context reload (#203150)](#203150) <!--- Backport version: 9.4.3 --> ### Questions ? Please refer to the [Backport tool documentation](https://github.com/sqren/backport) <!--BACKPORT not emitting\r\nthe up-to-date `data` when the parentApi search context updates. The\r\nretrieved `data` is stale and one step behind, so it is not reliable. cc\r\n@elastic/kibana-presentation\r\n\r\nAs @ThomThomson noticed in his test failure investigation another issue\r\nwas found in this PR due to the wrong handling of the searchSessionId\r\nwithin the Observability page (for more details see [his\r\nanalysis](https://github.com/elastic/kibana/pull/203150#issuecomment-2524080129)).\r\n@markov00 and @crespocarlos helped risolve this problem with some\r\nadditional changes on the Observability side of things: this will lead\r\nto some extra searchSessionId to be created, which will be eventually\r\nsolved by Observability team [shortly moving away from the\r\n`searchSessionId`\r\nmechanism](https://github.com/elastic/kibana/issues/203412)\r\n\r\n### Checklist\r\n\r\n- [x] [Unit or functional\r\ntests](https://www.elastic.co/guide/en/kibana/master/development-tests.html)\r\nwere updated or added to match the most common scenarios\r\n\r\n---------\r\n\r\nCo-authored-by: Marco Vettorello <marco.vettorello@elastic.co>\r\nCo-authored-by: Carlos Crespo <crespocarlos@users.noreply.github.com>","sha":"d4194ba5eb5a72960dad00cf956d9ea6e31219e0","branchLabelMapping":{"^v9.0.0$":"main","^v8.18.0$":"8.x","^v(\\d+).(\\d+).\\d+$":"$1.$2"}},"sourcePullRequest":{"labels":["release_note:fix","Team:Presentation","v9.0.0","backport:prev-minor","Feature:Embeddables","ci:project-deploy-observability","Team:obs-ux-management"],"title":"[Embeddable] Avoid rerendering loop due to search context reload","number":203150,"url":"https://github.com/elastic/kibana/pull/203150","mergeCommit":{"message":"[Embeddable] Avoid rerendering loop due to search context reload (#203150)\n\n## Summary\r\n\r\nFixes #202266\r\n\r\nThis PR fixes the underline rerendering issue at the `useSearchApi` hook\r\nlevel, so any embeddable component who uses this hook would benefit from\r\nthe fix.\r\n\r\nIdeally the props passed to the Lens component should be memoized, but\r\nthis assumption would break many integrations as the previous embeddable\r\ncomponent did take care of filtering duplicates.\r\nTo test this:\r\n* Go to `Observability > Alerts > Manage rules` and `Add Rule`, pick the\r\n`Custom threshold` option and verify the infinite reload does not happen\r\n\r\nOnce fixed this, another problem bubbled up with the brushing use case:\r\nwhen brushing a chart the chart was always one time range step behind.\r\nThe other bug was due to the `fetch$(api)` function propagating a stale\r\n`data` search context who the Lens embeddable was relying on.\r\nTo solve this other problem the following changes have been applied:\r\n* read the `searchSessionId` from the `api` directly (used for\r\n`autoRefresh`)\r\n* make sure to test the `Refresh` feature with both relative and\r\nabsolute time ranges\r\n* read the `search context` from the `parentApi` directly if implements\r\nthe `unifiedSearch` API\r\n* to test this, brush and check that the final time range matches the\r\ncorrect time range\r\n\r\n**Note**: the fundamental issue for the latter is `fetch# Backport This will backport the following commits from `main` to `8.x`: - [[Embeddable] Avoid rerendering loop due to search context reload (#203150)](#203150) <!--- Backport version: 9.4.3 --> ### Questions ? Please refer to the [Backport tool documentation](https://github.com/sqren/backport) <!--BACKPORT not emitting\r\nthe up-to-date `data` when the parentApi search context updates. The\r\nretrieved `data` is stale and one step behind, so it is not reliable. cc\r\n@elastic/kibana-presentation\r\n\r\nAs @ThomThomson noticed in his test failure investigation another issue\r\nwas found in this PR due to the wrong handling of the searchSessionId\r\nwithin the Observability page (for more details see [his\r\nanalysis](https://github.com/elastic/kibana/pull/203150#issuecomment-2524080129)).\r\n@markov00 and @crespocarlos helped risolve this problem with some\r\nadditional changes on the Observability side of things: this will lead\r\nto some extra searchSessionId to be created, which will be eventually\r\nsolved by Observability team [shortly moving away from the\r\n`searchSessionId`\r\nmechanism](https://github.com/elastic/kibana/issues/203412)\r\n\r\n### Checklist\r\n\r\n- [x] [Unit or functional\r\ntests](https://www.elastic.co/guide/en/kibana/master/development-tests.html)\r\nwere updated or added to match the most common scenarios\r\n\r\n---------\r\n\r\nCo-authored-by: Marco Vettorello <marco.vettorello@elastic.co>\r\nCo-authored-by: Carlos Crespo <crespocarlos@users.noreply.github.com>","sha":"d4194ba5eb5a72960dad00cf956d9ea6e31219e0"}},"sourceBranch":"main","suggestedTargetBranches":[],"targetPullRequestStates":[{"branch":"main","label":"v9.0.0","branchLabelMappingKey":"^v9.0.0$","isSourceBranch":true,"state":"MERGED","url":"https://github.com/elastic/kibana/pull/203150","number":203150,"mergeCommit":{"message":"[Embeddable] Avoid rerendering loop due to search context reload (#203150)\n\n## Summary\r\n\r\nFixes #202266\r\n\r\nThis PR fixes the underline rerendering issue at the `useSearchApi` hook\r\nlevel, so any embeddable component who uses this hook would benefit from\r\nthe fix.\r\n\r\nIdeally the props passed to the Lens component should be memoized, but\r\nthis assumption would break many integrations as the previous embeddable\r\ncomponent did take care of filtering duplicates.\r\nTo test this:\r\n* Go to `Observability > Alerts > Manage rules` and `Add Rule`, pick the\r\n`Custom threshold` option and verify the infinite reload does not happen\r\n\r\nOnce fixed this, another problem bubbled up with the brushing use case:\r\nwhen brushing a chart the chart was always one time range step behind.\r\nThe other bug was due to the `fetch$(api)` function propagating a stale\r\n`data` search context who the Lens embeddable was relying on.\r\nTo solve this other problem the following changes have been applied:\r\n* read the `searchSessionId` from the `api` directly (used for\r\n`autoRefresh`)\r\n* make sure to test the `Refresh` feature with both relative and\r\nabsolute time ranges\r\n* read the `search context` from the `parentApi` directly if implements\r\nthe `unifiedSearch` API\r\n* to test this, brush and check that the final time range matches the\r\ncorrect time range\r\n\r\n**Note**: the fundamental issue for the latter is `fetch# Backport This will backport the following commits from `main` to `8.x`: - [[Embeddable] Avoid rerendering loop due to search context reload (#203150)](#203150) <!--- Backport version: 9.4.3 --> ### Questions ? Please refer to the [Backport tool documentation](https://github.com/sqren/backport) <!--BACKPORT not emitting\r\nthe up-to-date `data` when the parentApi search context updates. The\r\nretrieved `data` is stale and one step behind, so it is not reliable. cc\r\n@elastic/kibana-presentation\r\n\r\nAs @ThomThomson noticed in his test failure investigation another issue\r\nwas found in this PR due to the wrong handling of the searchSessionId\r\nwithin the Observability page (for more details see [his\r\nanalysis](https://github.com/elastic/kibana/pull/203150#issuecomment-2524080129)).\r\n@markov00 and @crespocarlos helped risolve this problem with some\r\nadditional changes on the Observability side of things: this will lead\r\nto some extra searchSessionId to be created, which will be eventually\r\nsolved by Observability team [shortly moving away from the\r\n`searchSessionId`\r\nmechanism](https://github.com/elastic/kibana/issues/203412)\r\n\r\n### Checklist\r\n\r\n- [x] [Unit or functional\r\ntests](https://www.elastic.co/guide/en/kibana/master/development-tests.html)\r\nwere updated or added to match the most common scenarios\r\n\r\n---------\r\n\r\nCo-authored-by: Marco Vettorello <marco.vettorello@elastic.co>\r\nCo-authored-by: Carlos Crespo <crespocarlos@users.noreply.github.com>","sha":"d4194ba5eb5a72960dad00cf956d9ea6e31219e0"}}]}] BACKPORT--> Co-authored-by: Marco Liberati <dej611@users.noreply.github.com>
CAWilson94
pushed a commit
to CAWilson94/kibana
that referenced
this pull request
Dec 12, 2024
…stic#203150) ## Summary Fixes elastic#202266 This PR fixes the underline rerendering issue at the `useSearchApi` hook level, so any embeddable component who uses this hook would benefit from the fix. Ideally the props passed to the Lens component should be memoized, but this assumption would break many integrations as the previous embeddable component did take care of filtering duplicates. To test this: * Go to `Observability > Alerts > Manage rules` and `Add Rule`, pick the `Custom threshold` option and verify the infinite reload does not happen Once fixed this, another problem bubbled up with the brushing use case: when brushing a chart the chart was always one time range step behind. The other bug was due to the `fetch$(api)` function propagating a stale `data` search context who the Lens embeddable was relying on. To solve this other problem the following changes have been applied: * read the `searchSessionId` from the `api` directly (used for `autoRefresh`) * make sure to test the `Refresh` feature with both relative and absolute time ranges * read the `search context` from the `parentApi` directly if implements the `unifiedSearch` API * to test this, brush and check that the final time range matches the correct time range **Note**: the fundamental issue for the latter is `fetch$` not emitting the up-to-date `data` when the parentApi search context updates. The retrieved `data` is stale and one step behind, so it is not reliable. cc @elastic/kibana-presentation As @ThomThomson noticed in his test failure investigation another issue was found in this PR due to the wrong handling of the searchSessionId within the Observability page (for more details see [his analysis](elastic#203150 (comment))). @markov00 and @crespocarlos helped risolve this problem with some additional changes on the Observability side of things: this will lead to some extra searchSessionId to be created, which will be eventually solved by Observability team [shortly moving away from the `searchSessionId` mechanism](elastic#203412) ### Checklist - [x] [Unit or functional tests](https://www.elastic.co/guide/en/kibana/master/development-tests.html) were updated or added to match the most common scenarios --------- Co-authored-by: Marco Vettorello <marco.vettorello@elastic.co> Co-authored-by: Carlos Crespo <crespocarlos@users.noreply.github.com>
1 task
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
9 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.
Summary
Fixes #202266
This PR fixes the underline rerendering issue at the
useSearchApihook level, so any embeddable component who uses this hook would benefit from the fix.Ideally the props passed to the Lens component should be memoized, but this assumption would break many integrations as the previous embeddable component did take care of filtering duplicates.
To test this:
Observability > Alerts > Manage rulesandAdd Rule, pick theCustom thresholdoption and verify the infinite reload does not happenOnce fixed this, another problem bubbled up with the brushing use case: when brushing a chart the chart was always one time range step behind. The other bug was due to the
fetch$(api)function propagating a staledatasearch context who the Lens embeddable was relying on.To solve this other problem the following changes have been applied:
searchSessionIdfrom theapidirectly (used forautoRefresh)Refreshfeature with both relative and absolute time rangessearch contextfrom theparentApidirectly if implements theunifiedSearchAPINote: the fundamental issue for the latter is
fetch$not emitting the up-to-datedatawhen the parentApi search context updates. The retrieveddatais stale and one step behind, so it is not reliable. cc @elastic/kibana-presentationAs @ThomThomson noticed in his test failure investigation another issue was found in this PR due to the wrong handling of the searchSessionId within the Observability page (for more details see his analysis).
@markov00 and @crespocarlos helped risolve this problem with some additional changes on the Observability side of things: this will lead to some extra searchSessionId to be created, which will be eventually solved by Observability team shortly moving away from the
searchSessionIdmechanismChecklist