-
Notifications
You must be signed in to change notification settings - Fork 38.7k
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
Flake fix: validate each expression once in primary CEL correctness suite #109238
Flake fix: validate each expression once in primary CEL correctness suite #109238
Conversation
/priority critical-urgent |
We have test on expression cost but miss the test for hitting limit. Maybe worth to add a separate small test aside for limit check? |
/lgtm |
/assign @liggitt |
FYI: We have existing test covering runtime cost budget exceed in k8s.io/apiextensions-apiserver/pkg/apiserver/schema/defaulting/validation_test.go so removing the similar test from cel/validation_test should not be a big concern |
/lgtm #109239 is likely to resolve the tragically high integration test flake rate |
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: jpbetz, liggitt The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
mmm... that's specific to defaulting, though, right? that doesn't seem like the same coverage this is dropping We have test on expression cost but miss the test for hitting limit. Maybe worth to add a separate small test aside for limit check? that's a good point. a single test of the limit (rather than exercising N times on N expressions) seems important |
/milestone v1.24 |
(if the flake is bad, feel free to unhold and re-add that solo test next week) |
What type of PR is this?
/kind cleanup
/kind flake
What this PR does / why we need it:
The test suite is compiling and validating each of the many rules it tests three times. The 2nd two times are just to do a sanity check of the cost system, which isn't particularly valuable to test here given all the other tests that cover that.
I've reduced this to a single compile/validate call which reduces the CPU cost of the test by more than half.
Which issue(s) this PR fixes:
Fixes #109225
Special notes for your reviewer:
Does this PR introduce a user-facing change?
/sig api-machinery