[Fleet] Remove duplicative retries from client-side requests to APIs that depend on EPR #190722
Conversation
This file contains 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
kpollich
added
release_note:fix
Team:Fleet
|
Pinging @elastic/fleet (Team:Fleet) |
🤖 GitHub commentsExpand to view the GitHub comments
Just comment with:
|
nchaulet
reviewed
Aug 19, 2024
|
I moved to using Screen.Recording.2024-08-20.at.8.34.13.AM.mov |
|
@elasticmachine merge upstream |
💚 Build Succeeded
Metrics [docs]Async chunks
Page load bundle
History
To update your PR or re-run it, just comment with: cc @kpollich |
kibanamachine
pushed a commit
to kibanamachine/kibana
that referenced
this pull request
Aug 20, 2024
…that depend on EPR (elastic#190722) ## Summary Relates to elastic#136617 For APIs that depend on Fleet connecting to Elastic Package Registry, Fleet already retries the connections to EPR on the server side. This results in a situation where, when EPR is unreachable, the requests is retried several times on the server side, and then the request is retried again on the client-side by react-query. This results in very long running API requests. Since the server-side retries generally cover any kind of flakiness here, disabling the retry logic on the front-end seems sensible. ~I've also reduced the number of retries on the server side from 5 to 3 to help fail faster here.~ I walked back the retry change after some test failures, and I don't think it makes a big enough impact to justify changing. ## To test Set `xpack.fleet.registryUrl: 127.0.0.1:8080` with nothing running ## Before The requests spin for a very long time. https://github.com/user-attachments/assets/e4fd77ee-b36c-4965-9f71-e5b3e195f75e ## After The requests stop spinning after a few seconds as the retries won't loop. https://github.com/user-attachments/assets/82adc595-1bc4-4269-8501-2eb83525ad15 cc @shahzad31 (cherry picked from commit cf3149e)
💚 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 20, 2024
…o APIs that depend on EPR (#190722) (#190816) # Backport This will backport the following commits from `main` to `8.15`: - [[Fleet] Remove duplicative retries from client-side requests to APIs that depend on EPR (#190722)](#190722) <!--- Backport version: 9.4.3 --> ### Questions ? Please refer to the [Backport tool documentation](https://github.com/sqren/backport) <!--BACKPORT [{"author":{"name":"Kyle Pollich","email":"kyle.pollich@elastic.co"},"sourceCommit":{"committedDate":"2024-08-20T14:50:00Z","message":"[Fleet] Remove duplicative retries from client-side requests to APIs that depend on EPR (#190722)\n\n## Summary\r\n\r\nRelates to https://github.com/elastic/kibana/issues/136617\r\n\r\nFor APIs that depend on Fleet connecting to Elastic Package Registry,\r\nFleet already retries the connections to EPR on the server side. This\r\nresults in a situation where, when EPR is unreachable, the requests is\r\nretried several times on the server side, and then the request is\r\nretried again on the client-side by react-query. This results in very\r\nlong running API requests.\r\n\r\nSince the server-side retries generally cover any kind of flakiness\r\nhere, disabling the retry logic on the front-end seems sensible. ~I've\r\nalso reduced the number of retries on the server side from 5 to 3 to\r\nhelp fail faster here.~ I walked back the retry change after some test\r\nfailures, and I don't think it makes a big enough impact to justify\r\nchanging.\r\n\r\n## To test\r\n\r\nSet `xpack.fleet.registryUrl: 127.0.0.1:8080` with nothing running\r\n\r\n## Before\r\n\r\nThe requests spin for a very long time.\r\n\r\n\r\nhttps://github.com/user-attachments/assets/e4fd77ee-b36c-4965-9f71-e5b3e195f75e\r\n\r\n## After\r\n\r\nThe requests stop spinning after a few seconds as the retries won't\r\nloop.\r\n\r\n\r\nhttps://github.com/user-attachments/assets/82adc595-1bc4-4269-8501-2eb83525ad15\r\n\r\ncc @shahzad31","sha":"cf3149e983c5aec547e08cfa9202b68cd7115899","branchLabelMapping":{"^v8.16.0$":"main","^v(\\d+).(\\d+).\\d+$":"$1.$2"}},"sourcePullRequest":{"labels":["release_note:fix","Team:Fleet","backport:prev-minor","v8.16.0"],"title":"[Fleet] Remove duplicative retries from client-side requests to APIs that depend on EPR","number":190722,"url":"https://github.com/elastic/kibana/pull/190722","mergeCommit":{"message":"[Fleet] Remove duplicative retries from client-side requests to APIs that depend on EPR (#190722)\n\n## Summary\r\n\r\nRelates to https://github.com/elastic/kibana/issues/136617\r\n\r\nFor APIs that depend on Fleet connecting to Elastic Package Registry,\r\nFleet already retries the connections to EPR on the server side. This\r\nresults in a situation where, when EPR is unreachable, the requests is\r\nretried several times on the server side, and then the request is\r\nretried again on the client-side by react-query. This results in very\r\nlong running API requests.\r\n\r\nSince the server-side retries generally cover any kind of flakiness\r\nhere, disabling the retry logic on the front-end seems sensible. ~I've\r\nalso reduced the number of retries on the server side from 5 to 3 to\r\nhelp fail faster here.~ I walked back the retry change after some test\r\nfailures, and I don't think it makes a big enough impact to justify\r\nchanging.\r\n\r\n## To test\r\n\r\nSet `xpack.fleet.registryUrl: 127.0.0.1:8080` with nothing running\r\n\r\n## Before\r\n\r\nThe requests spin for a very long time.\r\n\r\n\r\nhttps://github.com/user-attachments/assets/e4fd77ee-b36c-4965-9f71-e5b3e195f75e\r\n\r\n## After\r\n\r\nThe requests stop spinning after a few seconds as the retries won't\r\nloop.\r\n\r\n\r\nhttps://github.com/user-attachments/assets/82adc595-1bc4-4269-8501-2eb83525ad15\r\n\r\ncc @shahzad31","sha":"cf3149e983c5aec547e08cfa9202b68cd7115899"}},"sourceBranch":"main","suggestedTargetBranches":[],"targetPullRequestStates":[{"branch":"main","label":"v8.16.0","branchLabelMappingKey":"^v8.16.0$","isSourceBranch":true,"state":"MERGED","url":"https://github.com/elastic/kibana/pull/190722","number":190722,"mergeCommit":{"message":"[Fleet] Remove duplicative retries from client-side requests to APIs that depend on EPR (#190722)\n\n## Summary\r\n\r\nRelates to https://github.com/elastic/kibana/issues/136617\r\n\r\nFor APIs that depend on Fleet connecting to Elastic Package Registry,\r\nFleet already retries the connections to EPR on the server side. This\r\nresults in a situation where, when EPR is unreachable, the requests is\r\nretried several times on the server side, and then the request is\r\nretried again on the client-side by react-query. This results in very\r\nlong running API requests.\r\n\r\nSince the server-side retries generally cover any kind of flakiness\r\nhere, disabling the retry logic on the front-end seems sensible. ~I've\r\nalso reduced the number of retries on the server side from 5 to 3 to\r\nhelp fail faster here.~ I walked back the retry change after some test\r\nfailures, and I don't think it makes a big enough impact to justify\r\nchanging.\r\n\r\n## To test\r\n\r\nSet `xpack.fleet.registryUrl: 127.0.0.1:8080` with nothing running\r\n\r\n## Before\r\n\r\nThe requests spin for a very long time.\r\n\r\n\r\nhttps://github.com/user-attachments/assets/e4fd77ee-b36c-4965-9f71-e5b3e195f75e\r\n\r\n## After\r\n\r\nThe requests stop spinning after a few seconds as the retries won't\r\nloop.\r\n\r\n\r\nhttps://github.com/user-attachments/assets/82adc595-1bc4-4269-8501-2eb83525ad15\r\n\r\ncc @shahzad31","sha":"cf3149e983c5aec547e08cfa9202b68cd7115899"}}]}] BACKPORT--> Co-authored-by: Kyle Pollich <kyle.pollich@elastic.co>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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
Relates to #136617
For APIs that depend on Fleet connecting to Elastic Package Registry, Fleet already retries the connections to EPR on the server side. This results in a situation where, when EPR is unreachable, the requests is retried several times on the server side, and then the request is retried again on the client-side by react-query. This results in very long running API requests.
Since the server-side retries generally cover any kind of flakiness here, disabling the retry logic on the front-end seems sensible.
I've also reduced the number of retries on the server side from 5 to 3 to help fail faster here.I walked back the retry change after some test failures, and I don't think it makes a big enough impact to justify changing.To test
Set
xpack.fleet.registryUrl: 127.0.0.1:8080with nothing runningBefore
The requests spin for a very long time.
Screen.Recording.2024-08-19.at.1.02.16.PM.mov
After
The requests stop spinning after a few seconds as the retries won't loop.
Screen.Recording.2024-08-19.at.12.54.38.PM.mov
cc @shahzad31