Handle InitialScale != 1 when creating multiScaler and some tidy-ups#8612
Handle InitialScale != 1 when creating multiScaler and some tidy-ups#8612knative-prow-robot merged 1 commit intoknative:masterfrom
Conversation
Now InitialScale has landed, take it into account when calculating initial EBC in multiscaler. Also: - Tidy up some comments. - Fix up some test error messages. - Move mutexes above the things they're guarding. - Use %d and %0.3f for printing numbers, consistently with existing Debug statement at start of same method.
knative-prow-robot
left a comment
There was a problem hiding this comment.
@julz: 0 warnings.
Details
In response to this:
Now InitialScale has landed, take it into account when calculating initial EBC in multiscaler.
Also:
- Tidy up some comments.
- Fix up test assertion messages.
- Move mutexes above the things they're guarding.
- Use %d and %0.3f for printing numbers, consistently with existing Debug statement at start of same method.
/assign @vagababov @markusthoemmes @taragu
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.
|
The following is the coverage report on the affected files.
|
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: julz, vagababov The full list of commands accepted by this bot can be found here. The pull request process is described here DetailsNeeds approval from an approver in each of these files:
Approvers can indicate their approval by writing |
alpha note added update checksum fix comments checksum and validation Update config/core/configmaps/gc.yaml Co-authored-by: Victor Agababov <vagababov@gmail.com> fix comments fix unit test fix both check include unit test make new settings for v2 v2 gc settings fix defaults default values fix unit test. expand comment fix validation error Widen collector test timeouts and some cleansing. (knative#8607) Make sorting consistent in GC test. (knative#8606) One of the tests ("keep recent lastPinned") failed fairly often because it used two revisions with the same LastPinned timestamp. Sorting a slice is unstable so in a lot of occasions, the given order was flipped and revision 5555 was tried to be deleted, even though it shouldn't even have been considered to be deleted. Note: sort.SliceStable doesn't help because the incoming list's order is already non-consistent. [master] Auto-update dependencies (knative#8611) Produced via: `./hack/update-deps.sh --upgrade && ./hack/update-codegen.sh` /assign dprotaso vagababov /cc dprotaso vagababov Add route consistency retry to new logging E2E test. (knative#8324) Handle InitialScale != 1 when creating multiScaler and some tidy-ups (knative#8612) Now InitialScale has landed, take it into account when calculating initial EBC in multiscaler. Also: - Tidy up some comments. - Fix up some test error messages. - Move mutexes above the things they're guarding. - Use %d and %0.3f for printing numbers, consistently with existing Debug statement at start of same method. Enable HA by default (knative#8602) * Enable HA by default * Centralize the bucket/replica information for running the HA testing. Add a test to ensure the count of the controller's reconciler stays in sync with the actual controller binary. Remove endpoints from KPA as well. (knative#8616) * Remove endpoints from KPA as well. Today I noticed that I missed the endpoints code in the KPA itself, when removing that informer. So now really get rid of those. The tests are of course nightmarish :) * remove old cruft Assorted fixes to enable chaos duck. (knative#8619) * Assorted fixes to enable chaos duck. This lands a handful of fixes that I uncovered preparing to run controlplane chaos testing during our e2e tests. * Drop sleep to 10s
* Create a maximum revision GC setting * update yaml alpha note added update checksum fix comments checksum and validation Update config/core/configmaps/gc.yaml Co-authored-by: Victor Agababov <vagababov@gmail.com> fix comments fix unit test fix both check include unit test make new settings for v2 v2 gc settings fix defaults default values fix unit test. expand comment fix validation error Widen collector test timeouts and some cleansing. (#8607) Make sorting consistent in GC test. (#8606) One of the tests ("keep recent lastPinned") failed fairly often because it used two revisions with the same LastPinned timestamp. Sorting a slice is unstable so in a lot of occasions, the given order was flipped and revision 5555 was tried to be deleted, even though it shouldn't even have been considered to be deleted. Note: sort.SliceStable doesn't help because the incoming list's order is already non-consistent. [master] Auto-update dependencies (#8611) Produced via: `./hack/update-deps.sh --upgrade && ./hack/update-codegen.sh` /assign dprotaso vagababov /cc dprotaso vagababov Add route consistency retry to new logging E2E test. (#8324) Handle InitialScale != 1 when creating multiScaler and some tidy-ups (#8612) Now InitialScale has landed, take it into account when calculating initial EBC in multiscaler. Also: - Tidy up some comments. - Fix up some test error messages. - Move mutexes above the things they're guarding. - Use %d and %0.3f for printing numbers, consistently with existing Debug statement at start of same method. Enable HA by default (#8602) * Enable HA by default * Centralize the bucket/replica information for running the HA testing. Add a test to ensure the count of the controller's reconciler stays in sync with the actual controller binary. Remove endpoints from KPA as well. (#8616) * Remove endpoints from KPA as well. Today I noticed that I missed the endpoints code in the KPA itself, when removing that informer. So now really get rid of those. The tests are of course nightmarish :) * remove old cruft Assorted fixes to enable chaos duck. (#8619) * Assorted fixes to enable chaos duck. This lands a handful of fixes that I uncovered preparing to run controlplane chaos testing during our e2e tests. * Drop sleep to 10s * Use the new GC settings * actually use the new settings * simplify * now it compiles * fix config references * test with min settings * max tests * change name of max to max-non-active * comments and validate positive * parse forever constant * fixing the unit test, better examples * include disabled setinel * merge in settings * disabled logic * update comments * review suggestions * fix const * update with logging changes * fix unit test * comment settings * unit test update * nits * fix unit test names * more consistent name * fix max boundary * nit * remove ref to lastpinned * review suggestions * fix sign * use nonactive as min * first filter out active * comment nit on config * swap if/else ordering * simplify * whitespace
Now InitialScale has landed, take it into account when calculating initial EBC in multiscaler.
Also:
/assign @vagababov @markusthoemmes @taragu