Stricter notion of esReacheable: require health response #5796
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
|
run/e2e-tests tags=es |
barkbay
reviewed
Jun 23, 2022
There was a problem hiding this comment.
Thanks for raising this PR, I agree that the current status may be confusing.
I'm must confess however I'm a bit hesitant regarding the way to improve things as there can be 2 reasons for having an "unknown" state:
The operator has not attempted to reach the cluster yet, in which case I would expect the condition to becorev1.ConditionUnknown. We don't really no if Elasticsearch is not responding to requests.The operator attempted to get the cluster state, but an error occurred (time out, credentials issue...). Health is still unknown but we now have an error to explain why. The condition should then becorev1.ConditionFalse, with maybe the error as the condition message.
Edit: Sorry I was focusing on the code while working on the PR and forgot your comment 🤦
That being said you can argue that the observer should attempt a first connection pretty quickly once the operator is started or when a cluster is created, so may be I'm overthinking...
barkbay
approved these changes
Jun 23, 2022
Co-authored-by: Michael Morello <[email protected]>
|
cla/check |
This was referenced Jan 4, 2023
fantapsody
pushed a commit
to fantapsody/cloud-on-k8s
that referenced
this pull request
Feb 7, 2023
The goal is avoid confusing error messages like "could not verify license" but instead have a meaningful status like "Elasticsearch is not reachable". The implication here is that esReachable is now tied to the first successful observation of cluster health. Those happen asynchronously in the observer mechanism and are not done inside the reconciliation loop. I think that is OK as the cluster is not available immediately but with a certain delay anyway, depending on hardware and the time it takes for the ES container in the Pods to become ready. If the timing is unfortunate we might lose an additional 10 seconds (which is the current observation interval) which seems acceptable to me. Co-authored-by: Michael Morello <[email protected]>
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.
Fixes #5776
Simplest (?) possible fix without doing extra requests. The goal is avoid confusing error messages like "could not verify license" but instead have a meaningful status like "Elasticsearch is not reachable". The implication here is that
esReachableis now tied to the first successful observation of cluster health. Those happen asynchronously in the observer mechanism and are not done inside the reconciliation loop. I think that is OK as the cluster is not available immediately but with a certain delay anyway, depending on hardware and the time it takes for the ES container in the Pods to become ready. If the timing is unfortunate we might lose an additional 10 seconds (which is the current observation interval) which seems acceptable to me.