[Fleet] Fix agent policy selection and creation when there are a large number of agent policies ( > 1k) #151119
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
hop-dev
added
release_note:fix
Team:Fleet
💚 Build Succeeded
Metrics [docs]Async chunks
Page load bundle
To update your PR or re-run it, just comment with: cc @hop-dev |
|
Pinging @elastic/fleet (Team:Fleet) |
| - schema: | ||
| type: boolean | ||
| in: query | ||
| name: noAgentCount |
There was a problem hiding this comment.
I would find it more intuitive to have a flag when we want to include agentCount, rather than having a flag to disable. Though this might be a bigger refactor if the API is already used in many places.
There was a problem hiding this comment.
I agree, I was so tempted to do it but this change is already larger than I expected. It would be a breaking change to the API
There was a problem hiding this comment.
You're right, we shouldn't do a breaking change.
juliaElastic
approved these changes
Feb 14, 2023
kibanamachine
pushed a commit
to kibanamachine/kibana
that referenced
this pull request
Feb 14, 2023
…e number of agent policies ( > 1k) (elastic#151119) ## Summary Closes elastic#150605 When adding an integration, the agent policy selector (create or select existing agent policy) does not work if there are too many agent policies. In an environment with a large number of agent policies (e.g 1000+) the following API request times out or takes 30 - 60 seconds to respond: ``` http://localhost:5601/mark/api/fleet/agent_policies?page=1&perPage=10000&sortField=name&sortOrder=asc&full=true ``` That is because for each agent policy we get the saved object twice, get ALL package policies, and count the agents, so for 1000 agents we fun 4000 queries using `pmap` in some places. this PR changes it so that we do not get the agent count or `full` agent policy for every agent policy when populating the agent policy select, instead we only get the selected agent policy. This has meant that I have to separately query package policies in order to do the limited packages and APM output checks, but the component now loads instantly on an environment with 3k agent policies. Key changes: - add `noAgentCount` query param to agent policies API, to not count the agents for all agent policies - populate `agents` in the get one agent policy API, this was an inconsistency in our agent policy API - when selecting an existing agent policy, only get the full agent policy for the selected agent policy ###Testing No behaviour should have changed for the agent policy select, some niche behaviour of the selector: - if adding the APM integration, any agent policy with a logstash output configured for integration data should be disabled - if adding a limited integration e.g apm or defend, any agent policy already containing that integration should be disabled <img width="1052" alt="Screenshot 2023-02-14 at 10 45 03" src="https://user-images.githubusercontent.com/3315046/218745430-4a8f0ded-1e0b-4319-bc2c-cc5253b4cdd2.png"> (cherry picked from commit 9a52ef4)
💚 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
Feb 14, 2023
…a large number of agent policies ( > 1k) (#151119) (#151131) # Backport This will backport the following commits from `main` to `8.7`: - [[Fleet] Fix agent policy selection and creation when there are a large number of agent policies ( > 1k) (#151119)](#151119) <!--- Backport version: 8.9.7 --> ### Questions ? Please refer to the [Backport tool documentation](https://github.com/sqren/backport) <!--BACKPORT [{"author":{"name":"Mark Hopkin","email":"[email protected]"},"sourceCommit":{"committedDate":"2023-02-14T13:36:17Z","message":"[Fleet] Fix agent policy selection and creation when there are a large number of agent policies ( > 1k) (#151119)\n\n## Summary\r\n\r\nCloses #150605\r\n\r\nWhen adding an integration, the agent policy selector (create or select\r\nexisting agent policy) does not work if there are too many agent\r\npolicies.\r\n\r\nIn an environment with a large number of agent policies (e.g 1000+) the\r\nfollowing API request times out or takes 30 - 60 seconds to respond:\r\n\r\n```\r\nhttp://localhost:5601/mark/api/fleet/agent_policies?page=1&perPage=10000&sortField=name&sortOrder=asc&full=true\r\n```\r\n\r\nThat is because for each agent policy we get the saved object twice, get\r\nALL package policies, and count the agents, so for 1000 agents we fun\r\n4000 queries using `pmap` in some places.\r\n\r\nthis PR changes it so that we do not get the agent count or `full` agent\r\npolicy for every agent policy when populating the agent policy select,\r\ninstead we only get the selected agent policy.\r\n\r\nThis has meant that I have to separately query package policies in order\r\nto do the limited packages and APM output checks, but the component now\r\nloads instantly on an environment with 3k agent policies.\r\n\r\nKey changes:\r\n\r\n- add `noAgentCount` query param to agent policies API, to not count the\r\nagents for all agent policies\r\n- populate `agents` in the get one agent policy API, this was an\r\ninconsistency in our agent policy API\r\n- when selecting an existing agent policy, only get the full agent\r\npolicy for the selected agent policy\r\n\r\n\r\n###Testing\r\n\r\nNo behaviour should have changed for the agent policy select, some niche\r\nbehaviour of the selector:\r\n\r\n- if adding the APM integration, any agent policy with a logstash output\r\nconfigured for integration data should be disabled\r\n- if adding a limited integration e.g apm or defend, any agent policy\r\nalready containing that integration should be disabled\r\n\r\n<img width=\"1052\" alt=\"Screenshot 2023-02-14 at 10 45 03\"\r\nsrc=\"https://user-images.githubusercontent.com/3315046/218745430-4a8f0ded-1e0b-4319-bc2c-cc5253b4cdd2.png\">","sha":"9a52ef4baf07baa8a877252e3a740cfde1624063","branchLabelMapping":{"^v8.8.0$":"main","^v(\\d+).(\\d+).\\d+$":"$1.$2"}},"sourcePullRequest":{"labels":["release_note:fix","Team:Fleet","v8.7.0","v8.8.0"],"number":151119,"url":"https://github.com/elastic/kibana/pull/151119","mergeCommit":{"message":"[Fleet] Fix agent policy selection and creation when there are a large number of agent policies ( > 1k) (#151119)\n\n## Summary\r\n\r\nCloses #150605\r\n\r\nWhen adding an integration, the agent policy selector (create or select\r\nexisting agent policy) does not work if there are too many agent\r\npolicies.\r\n\r\nIn an environment with a large number of agent policies (e.g 1000+) the\r\nfollowing API request times out or takes 30 - 60 seconds to respond:\r\n\r\n```\r\nhttp://localhost:5601/mark/api/fleet/agent_policies?page=1&perPage=10000&sortField=name&sortOrder=asc&full=true\r\n```\r\n\r\nThat is because for each agent policy we get the saved object twice, get\r\nALL package policies, and count the agents, so for 1000 agents we fun\r\n4000 queries using `pmap` in some places.\r\n\r\nthis PR changes it so that we do not get the agent count or `full` agent\r\npolicy for every agent policy when populating the agent policy select,\r\ninstead we only get the selected agent policy.\r\n\r\nThis has meant that I have to separately query package policies in order\r\nto do the limited packages and APM output checks, but the component now\r\nloads instantly on an environment with 3k agent policies.\r\n\r\nKey changes:\r\n\r\n- add `noAgentCount` query param to agent policies API, to not count the\r\nagents for all agent policies\r\n- populate `agents` in the get one agent policy API, this was an\r\ninconsistency in our agent policy API\r\n- when selecting an existing agent policy, only get the full agent\r\npolicy for the selected agent policy\r\n\r\n\r\n###Testing\r\n\r\nNo behaviour should have changed for the agent policy select, some niche\r\nbehaviour of the selector:\r\n\r\n- if adding the APM integration, any agent policy with a logstash output\r\nconfigured for integration data should be disabled\r\n- if adding a limited integration e.g apm or defend, any agent policy\r\nalready containing that integration should be disabled\r\n\r\n<img width=\"1052\" alt=\"Screenshot 2023-02-14 at 10 45 03\"\r\nsrc=\"https://user-images.githubusercontent.com/3315046/218745430-4a8f0ded-1e0b-4319-bc2c-cc5253b4cdd2.png\">","sha":"9a52ef4baf07baa8a877252e3a740cfde1624063"}},"sourceBranch":"main","suggestedTargetBranches":["8.7"],"targetPullRequestStates":[{"branch":"8.7","label":"v8.7.0","labelRegex":"^v(\\d+).(\\d+).\\d+$","isSourceBranch":false,"state":"NOT_CREATED"},{"branch":"main","label":"v8.8.0","labelRegex":"^v8.8.0$","isSourceBranch":true,"state":"MERGED","url":"https://github.com/elastic/kibana/pull/151119","number":151119,"mergeCommit":{"message":"[Fleet] Fix agent policy selection and creation when there are a large number of agent policies ( > 1k) (#151119)\n\n## Summary\r\n\r\nCloses #150605\r\n\r\nWhen adding an integration, the agent policy selector (create or select\r\nexisting agent policy) does not work if there are too many agent\r\npolicies.\r\n\r\nIn an environment with a large number of agent policies (e.g 1000+) the\r\nfollowing API request times out or takes 30 - 60 seconds to respond:\r\n\r\n```\r\nhttp://localhost:5601/mark/api/fleet/agent_policies?page=1&perPage=10000&sortField=name&sortOrder=asc&full=true\r\n```\r\n\r\nThat is because for each agent policy we get the saved object twice, get\r\nALL package policies, and count the agents, so for 1000 agents we fun\r\n4000 queries using `pmap` in some places.\r\n\r\nthis PR changes it so that we do not get the agent count or `full` agent\r\npolicy for every agent policy when populating the agent policy select,\r\ninstead we only get the selected agent policy.\r\n\r\nThis has meant that I have to separately query package policies in order\r\nto do the limited packages and APM output checks, but the component now\r\nloads instantly on an environment with 3k agent policies.\r\n\r\nKey changes:\r\n\r\n- add `noAgentCount` query param to agent policies API, to not count the\r\nagents for all agent policies\r\n- populate `agents` in the get one agent policy API, this was an\r\ninconsistency in our agent policy API\r\n- when selecting an existing agent policy, only get the full agent\r\npolicy for the selected agent policy\r\n\r\n\r\n###Testing\r\n\r\nNo behaviour should have changed for the agent policy select, some niche\r\nbehaviour of the selector:\r\n\r\n- if adding the APM integration, any agent policy with a logstash output\r\nconfigured for integration data should be disabled\r\n- if adding a limited integration e.g apm or defend, any agent policy\r\nalready containing that integration should be disabled\r\n\r\n<img width=\"1052\" alt=\"Screenshot 2023-02-14 at 10 45 03\"\r\nsrc=\"https://user-images.githubusercontent.com/3315046/218745430-4a8f0ded-1e0b-4319-bc2c-cc5253b4cdd2.png\">","sha":"9a52ef4baf07baa8a877252e3a740cfde1624063"}}]}] BACKPORT--> Co-authored-by: Mark Hopkin <[email protected]>
justinkambic
pushed a commit
to justinkambic/kibana
that referenced
this pull request
Feb 23, 2023
![https://user-images.githubusercontent.com/3315046/218745430-4a8f0ded-1e0b-4319-bc2c-cc5253b4cdd2.png\">","sha":"9a52ef4baf07baa8a877252e3a740cfde1624063","branchLabelMapping":{"^v8.8.0$":"main","^v(\\d+).(\\d+).\\d+$":"$1.$2"}},"sourcePullRequest":{"labels":["release_note:fix","Team:Fleet","v8.7.0","v8.8.0"],"number":151119,"url":"https://github.com/elastic/kibana/pull/151119","mergeCommit":{"message":"[Fleet]](https://user-images.githubusercontent.com/3315046/218745430-4a8f0ded-1e0b-4319-bc2c-cc5253b4cdd2.png\">","sha":"9a52ef4baf07baa8a877252e3a740cfde1624063","branchLabelMapping":{"^v8.8.0$":"main","^v(\\d+).(\\d+).\\d+$":"$1.$2"}},"sourcePullRequest":{"labels":["release_note:fix","Team:Fleet","v8.7.0","v8.8.0"],"number":151119,"url":"https://github.com/elastic/kibana/pull/151119","mergeCommit":{"message":"[Fleet]){kind=link}
![https://user-images.githubusercontent.com/3315046/218745430-4a8f0ded-1e0b-4319-bc2c-cc5253b4cdd2.png\">","sha":"9a52ef4baf07baa8a877252e3a740cfde1624063"}},"sourceBranch":"main","suggestedTargetBranches":["8.7"],"targetPullRequestStates":[{"branch":"8.7","label":"v8.7.0","labelRegex":"^v(\\d+).(\\d+).\\d+$","isSourceBranch":false,"state":"NOT_CREATED"},{"branch":"main","label":"v8.8.0","labelRegex":"^v8.8.0$","isSourceBranch":true,"state":"MERGED","url":"https://github.com/elastic/kibana/pull/151119","number":151119,"mergeCommit":{"message":"[Fleet]](https://user-images.githubusercontent.com/3315046/218745430-4a8f0ded-1e0b-4319-bc2c-cc5253b4cdd2.png\">","sha":"9a52ef4baf07baa8a877252e3a740cfde1624063"}},"sourceBranch":"main","suggestedTargetBranches":["8.7"],"targetPullRequestStates":[{"branch":"8.7","label":"v8.7.0","labelRegex":"^v(\\d+).(\\d+).\\d+$","isSourceBranch":false,"state":"NOT_CREATED"},{"branch":"main","label":"v8.8.0","labelRegex":"^v8.8.0$","isSourceBranch":true,"state":"MERGED","url":"https://github.com/elastic/kibana/pull/151119","number":151119,"mergeCommit":{"message":"[Fleet]){kind=link}
![https://user-images.githubusercontent.com/3315046/218745430-4a8f0ded-1e0b-4319-bc2c-cc5253b4cdd2.png\">","sha":"9a52ef4baf07baa8a877252e3a740cfde1624063](https://user-images.githubusercontent.com/3315046/218745430-4a8f0ded-1e0b-4319-bc2c-cc5253b4cdd2.png\">","sha":"9a52ef4baf07baa8a877252e3a740cfde1624063){kind=link}
…e number of agent policies ( > 1k) (elastic#151119) ## Summary Closes elastic#150605 When adding an integration, the agent policy selector (create or select existing agent policy) does not work if there are too many agent policies. In an environment with a large number of agent policies (e.g 1000+) the following API request times out or takes 30 - 60 seconds to respond: ``` http://localhost:5601/mark/api/fleet/agent_policies?page=1&perPage=10000&sortField=name&sortOrder=asc&full=true ``` That is because for each agent policy we get the saved object twice, get ALL package policies, and count the agents, so for 1000 agents we fun 4000 queries using `pmap` in some places. this PR changes it so that we do not get the agent count or `full` agent policy for every agent policy when populating the agent policy select, instead we only get the selected agent policy. This has meant that I have to separately query package policies in order to do the limited packages and APM output checks, but the component now loads instantly on an environment with 3k agent policies. Key changes: - add `noAgentCount` query param to agent policies API, to not count the agents for all agent policies - populate `agents` in the get one agent policy API, this was an inconsistency in our agent policy API - when selecting an existing agent policy, only get the full agent policy for the selected agent policy ###Testing No behaviour should have changed for the agent policy select, some niche behaviour of the selector: - if adding the APM integration, any agent policy with a logstash output configured for integration data should be disabled - if adding a limited integration e.g apm or defend, any agent policy already containing that integration should be disabled <img width="1052" alt="Screenshot 2023-02-14 at 10 45 03" src="https://user-images.githubusercontent.com/3315046/218745430-4a8f0ded-1e0b-4319-bc2c-cc5253b4cdd2.png">
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
Closes #150605
When adding an integration, the agent policy selector (create or select existing agent policy) does not work if there are too many agent policies.
In an environment with a large number of agent policies (e.g 1000+) the following API request times out or takes 30 - 60 seconds to respond:
That is because for each agent policy we get the saved object twice, get ALL package policies, and count the agents, so for 1000 agents we fun 4000 queries using
pmapin some places.this PR changes it so that we do not get the agent count or
fullagent policy for every agent policy when populating the agent policy select, instead we only get the selected agent policy.This has meant that I have to separately query package policies in order to do the limited packages and APM output checks, but the component now loads instantly on an environment with 3k agent policies.
Key changes:
noAgentCountquery param to agent policies API, to not count the agents for all agent policiesagentsin the get one agent policy API, this was an inconsistency in our agent policy API###Testing
No behaviour should have changed for the agent policy select, some niche behaviour of the selector: