Make IlmHealthIndicatorService work in multi-project cluster
#129961
Conversation
This does _not_ make the health indicator project-aware, it merely avoids exceptions in case there are multiple projects in the cluster. The health indicator would require significant refactoring to be made project-aware, which is not worth it since ILM will not be running in a multi-project context (i.e. serverless).
nielsbauman
added
>non-issue
:Data Management/ILM+SLM
elasticsearchmachine
added
v9.1.0
Team:Data Management
|
Pinging @elastic/es-data-management (Team:Data Management) |
| * This method solely exists because we are not making ILM properly project-aware and it's not worth the investment of altering this | ||
| * health indicator to be project-aware. | ||
| */ | ||
| private static ProjectMetadata getDefaultILMProject(ClusterState state) { |
There was a problem hiding this comment.
Is this going to get your new 'not MP compatible' annotation?
I assume that adding an assertion runs into the same problems that we have in the other cases?
There was a problem hiding this comment.
Yeah, I just added the annotation.
I assume that adding an assertion runs into the same problems that we have in the other cases?
Adding the assertion would lead to similar problems. Since this is not a cluster state listener, the impact is far less, but we still call the health report here. Even though it's just one test/test suite, I'm still not very comfortable with the idea of trying to modify the big XPackRestIT or this test (we could retrieve a specific indicator rather than the full health report). The former is a larger effort and the latter reduces test coverage IMO.
…ic#129961) This does _not_ make the health indicator project-aware, it merely avoids exceptions in case there are multiple projects in the cluster. The health indicator would require significant refactoring to be made project-aware, which is not worth it since ILM will not be running in a multi-project context (i.e. serverless).
This does not make the health indicator project-aware, it merely avoids exceptions in case there are multiple projects in the cluster. The health indicator would require significant refactoring to be made project-aware, which is not worth it since ILM will not be running in a multi-project context (i.e. serverless).