-
Notifications
You must be signed in to change notification settings - Fork 55
fix: skip NodeEvaluation upsert when evaluation did not run #218
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Open
sahitya-chandra
wants to merge
2
commits into
kubernetes-sigs:main
Choose a base branch
from
sahitya-chandra:fix/empty-nodeevaluation-status-patch
base: main
Could not load branches
Branch not found: {{ refName }}
Loading
Could not load tags
Nothing to show
Loading
Are you sure you want to change the base?
Some commits from the old base branch may be removed from the timeline,
and old review comments may become outdated.
+249
−25
Open
Changes from all commits
Commits
Show all changes
2 commits
Select commit
Hold shift + click to select a range
File filter
Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
There are no files selected for viewing
This file contains hidden or 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
This file contains hidden or 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
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Outdated suggestions cannot be applied.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.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I view this more of a UX improvement than a bug fix --
evaluateRuleForNodefails for a node, it patches the rule status by overriding with NodeEvaluation{}, erasing previous (if existing) eval result. This change prevents from doing it, but the risk is possibly leaving a stale result for that rule.We have discussed moving the status object to a separate 'NodeReadinessEvaluation' CRD (in v1alpha2) to capture per rule results. This flow could look like --
instead of double looping and x-updates we are currently doing here.
Assuming we move to a 1--1 evaluation result present for each Node evaluated by NodeReadinessController, how would the rule-evaluation be structured? and how should we record per rule failures in it?
@Karthik-K-N I think we should prioritize evaluating this to establish a cohesive observability for NRC ideally aligning it earlier with the mentorship timeline as there are some observability focus in it.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If we have multiple rules per node and one among the rule evaluation fails we capture them in the ReadinessReport: https://github.com/kubernetes-sigs/node-readiness-controller/pull/133/changes#r2857725522
Uh oh!
There was an error while loading. Please reload this page.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks @ajaysundark and @Karthik-K-N, I just went through #133, and the
NodeReadinessRuleReport(nrrp) CRD makes sense to me. I agree a per-node report keyed byruleNameis cleaner for these results than theNodeEvaluations[]list inside the rule status, so I'm happy to align with that.One small thing on the "UX or bug" point: I think it is a bit more than UX today. On the failure path, when the node has no prior cache
NodeEvaluation(i.e. first evaluation),currEvalis the zero value, so the patch carriesNodeEvaluation{}. Its required fields areomitempty, so it serializes to{}and the apiserver rejects the wholeStatus().Patchwith a 422.failedNodesis in that same patch, so it gets dropped too. So on that path the operator sees nothing, not evenfailedNodes, just a log line.I checked this with envtest against the apiserver: on
mainthe failed node ends withfailedNodesempty, and with this change it persists. So I feelfailedNodesonly really captures the failure once we stop including the empty eval in the patch. They still go out in the same Status().Patch() call, but without the invalid {} entry the 422 no longer firesOn the stale result point, I agree. The idea here is just to keep the last good eval instead of erasing it, but I'm fine dropping the eval on failure instead if you prefer
Either way, I'm happy to keep this as a small v1alpha1 fix and leave the bigger change to the report work :)