[Fleet] Fix managed agent policy preconfiguration update

[nchaulet]

Summary

Managed agent policy defined in the kibana config should be updated, if there is a change in the kibana config. It seems if the saved object for that agent policy is out-of-sync and do not have the is_managed attribute correctly set we never update them. (This is causing some bug like with the keep_monitoring_alive flag added in 8.12).

That PR fix that by using the policy defined in the config and not the SO as source of truth.

Tests

Automated

I added unit test to the preconfiguration service

Manual

You can start kibana locally with the following config

xpack.fleet.agentPolicies:
  - name: Test policy
    description: TEST
    id: test-policy
    is_default: false
    is_default_fleet_server: false
    package_policies: []

Than update it to add is_managed and keep_monitoring_alive and verify your policy contains these fields

xpack.fleet.agentPolicies:
  - name: Test policy
    description: TEST
    id: test-policy
    is_default: false
    is_managed: true
    is_default_fleet_server: false
    package_policies: []
    keep_monitoring_alive: true

Questions

Should we add a release note for that?

[apmmachine]

🤖 GitHub comments

Expand to view the GitHub comments

Just comment with:

• /oblt-deploy : Deploy a Kibana instance using the Observability test environments.
• run elasticsearch-ci/docs : Re-trigger the docs validation. (use unformatted text in the comment!)

[nchaulet]

/ci

[kibana-ci]

💛 Build succeeded, but was flaky

• Buildkite Build
• Commit: 0b478d1

Failed CI Steps

• Jest Integration Tests #2

Test Failures

• [job] [logs] Jest Integration Tests #2 / Migration actions - serverless environment readWithPit returns a left es_response_too_large error when a read batch exceeds the maxResponseSize

Metrics [docs]

✅ unchanged

To update your PR or re-run it, just comment with:
@elasticmachine merge upstream

cc @nchaulet

[elasticmachine]

Pinging @elastic/fleet (Team:Fleet)

[jen-huang]

I think it would be good to flag this as a fix in release notes.

[juliaElastic]

Is the keep_monitoring_alive supposed to only work for managed policies? I'm wondering if we should allow it for other, non-managed policies too.

[nchaulet]

Is the keep_monitoring_alive supposed to only work for managed policies? I'm wondering if we should allow it for other, non-managed policies too.

@juliaElastic I think nothing should prevent to setkeep_monitoring_alive for non managed policy it's available through the API, I am wondering what will be the use case but it should work.

[kibanamachine]

💚 All backports created successfully

Status	Branch	Result
✅	8.14

Note: Successful backport PRs will be merged automatically after passing CI.

Questions ?

Please refer to the Backport tool documentation

[nchaulet]

💚 All backports created successfully

Status	Branch	Result
✅	8.13

Note: Successful backport PRs will be merged automatically after passing CI.

Questions ?

Please refer to the Backport tool documentation