Bug 1957991: install/0000_90_cluster-version-operator_02_servicemonitor: Soften ClusterOperatorDegraded #554
Conversation
…usterOperatorDegraded During install, the CVO has pushed manifests into the cluster as fast as possible without blocking on "has the in-cluster resource leveled?" since way back in b0b4902 (clusteroperator: Don't block on failing during initialization, 2019-03-11, openshift#136). That can lead to ClusterOperatorDown and ClusterOperatorDegraded firing during install, as we see in [1], where: * ClusterOperatorDegraded started pending at 5:00:15Z [2]. * Install completed at 5:09:58Z [3]. * ClusterOperatorDegraded started firing at 5:10:04Z [2]. * ClusterOperatorDegraded stopped firing at 5:10:23Z [2]. * The e2e suite complained about [1]: alert ClusterOperatorDegraded fired for 15 seconds with labels: {... name="authentication"...} (open bug: https://bugzilla.redhat.com/show_bug.cgi?id=1939580) ClusterOperatorDown is similar, but I'll leave addressing it to a separate commit. For ClusterOperatorDegraded, the degraded condition should not be particularly urgent [4], so we should be find bumping it to 'warning' and using 'for: 30m' or something more relaxed than the current 10m. [1]: https://prow.ci.openshift.org/view/gs/origin-ci-test/logs/release-openshift-ocp-installer-e2e-aws-upi-4.8/1389436726862155776 [2]: https://promecieus.dptools.openshift.org/?search=https://prow.ci.openshift.org/view/gs/origin-ci-test/logs/release-openshift-ocp-installer-e2e-aws-upi-4.8/1389436726862155776 group by (alertstate) (ALERTS{alertname="ClusterOperatorDegraded"}) [3]: https://gcsweb-ci.apps.ci.l2s4.p1.openshiftapps.com/gcs/origin-ci-test/logs/release-openshift-ocp-installer-e2e-aws-upi-4.8/1389436726862155776/artifacts/e2e-aws-upi/clusterversion.json [4]: openshift/api#916
openshift-ci
bot
added
bugzilla/severity-medium
|
@wking: This pull request references Bugzilla bug 1957991, which is valid. The bug has been moved to the POST state. The bug has been updated to refer to the pull request using the external bug tracker. 3 validation(s) were run on this bug
Requesting review from QA contact: DetailsIn response to this:
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. |
openshift-ci
bot
added
the
approved
|
@wking: This pull request references Bugzilla bug 1957991, which is valid. 3 validation(s) were run on this bug
Requesting review from QA contact: DetailsIn response to this:
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. |
| for: 30m | ||
| labels: | ||
| severity: critical | ||
| severity: warning |
There was a problem hiding this comment.
I’m ok with this. Warnings and degraded are for the morning. Do we have an estimate of how much churn this removes in the fleet?
There was a problem hiding this comment.
Poking around in recent Telemetry, the improvement for running clusters is expected to be small but noticeable. I expect impact during install to be larger, but haven't worked up numbers. We should be able to scrape it out of test-case JUnit post-merge.
|
/approve Has this mindset change been communicated to admin console and support teams? |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: smarterclayton, wking 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 |
|
/retest Please review the full test history for this PR and help us cut down flakes. |
6 similar comments
|
/retest Please review the full test history for this PR and help us cut down flakes. |
|
/retest Please review the full test history for this PR and help us cut down flakes. |
|
/retest Please review the full test history for this PR and help us cut down flakes. |
|
/retest Please review the full test history for this PR and help us cut down flakes. |
|
/retest Please review the full test history for this PR and help us cut down flakes. |
|
/retest Please review the full test history for this PR and help us cut down flakes. |
|
CC @openshift/sre-alert-sme, in case you want to review the changes. |
|
upgrade failure is unrelated Kube-API-server connectivity issues on Azure. Shouldn't block this landing. /override ci/prow/e2e-agnostic-upgrade |
|
@wking: Overrode contexts on behalf of wking: ci/prow/e2e-agnostic-upgrade DetailsIn response to this:
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. |
|
@wking: All pull requests linked via external trackers have merged: Bugzilla bug 1957991 has been moved to the MODIFIED state. DetailsIn response to this:
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. |
| group by (name) (cluster_operator_up{job="cluster-version-operator"}) | ||
| ) == 1 | ||
| for: 10m | ||
| for: 30m |
There was a problem hiding this comment.
The 83 line has such message - "Cluster operator {{ "{{ $labels.name }}" }} has been degraded for 10 minutes". Should be also update "10 minutes" to "30 minutes"?
…usterOperatorDegraded message to 30m Catching up with fb5257d (install/0000_90_cluster-version-operator_02_servicemonitor: Soften ClusterOperatorDegraded, 2021-05-06, openshift#554).
…usterOperatorDegraded During install, the CVO has pushed manifests into the cluster as fast as possible without blocking on "has the in-cluster resource leveled?" since way back in b0b4902 (clusteroperator: Don't block on failing during initialization, 2019-03-11, openshift#136). That can lead to ClusterOperatorDown and ClusterOperatorDegraded firing during install, as we see in [1], where: * ClusterOperatorDegraded started pending at 5:00:15Z [2]. * Install completed at 5:09:58Z [3]. * ClusterOperatorDegraded started firing at 5:10:04Z [2]. * ClusterOperatorDegraded stopped firing at 5:10:23Z [2]. * The e2e suite complained about [1]: alert ClusterOperatorDegraded fired for 15 seconds with labels: {... name="authentication"...} (open bug: https://bugzilla.redhat.com/show_bug.cgi?id=1939580) ClusterOperatorDown is similar, but I'll leave addressing it to a separate commit. For ClusterOperatorDegraded, the degraded condition should not be particularly urgent [4], so we should be find bumping it to 'warning' and using 'for: 30m' or something more relaxed than the current 10m. This commit brings back * fb5257d (install/0000_90_cluster-version-operator_02_servicemonitor: Soften ClusterOperatorDegraded, 2021-05-06, openshift#554) and * 92ed7f1 (install/0000_90_cluster-version-operator_02_servicemonitor: Update ClusterOperatorDegraded message to 30m, 2021-05-08, openshift#556). There are some conflicts, because I am not bringing back 90539f9 (pkg/cvo/metrics: Ignore Degraded for cluster_operator_up, 2021-04-26, openshift#550). But that one had its own conflicts in metrics.go [5], and the conflicts with this commit were orthogonal context issues, so moving this back to 4.7 first won't make it much harder to bring back openshift#550 and such later on, if we decide to do that. [1]: https://prow.ci.openshift.org/view/gs/origin-ci-test/logs/release-openshift-ocp-installer-e2e-aws-upi-4.8/1389436726862155776 [2]: https://promecieus.dptools.openshift.org/?search=https://prow.ci.openshift.org/view/gs/origin-ci-test/logs/release-openshift-ocp-installer-e2e-aws-upi-4.8/1389436726862155776 group by (alertstate) (ALERTS{alertname="ClusterOperatorDegraded"}) [3]: https://gcsweb-ci.apps.ci.l2s4.p1.openshiftapps.com/gcs/origin-ci-test/logs/release-openshift-ocp-installer-e2e-aws-upi-4.8/1389436726862155776/artifacts/e2e-aws-upi/clusterversion.json [4]: openshift/api#916 [5]: openshift#550 (comment)
6 participants
During install, the CVO has pushed manifests into the cluster as fast as possible without blocking on "has the in-cluster resource leveled?" since way back in b0b4902 (#136). That can lead to
ClusterOperatorDownandClusterOperatorDegradedfiring during install, as we see here, where:ClusterOperatorDegraded started pending at 5:00:15Z (
group by (alertstate) (ALERTS{alertname="ClusterOperatorDegraded"}).Install completed at 5:09:58Z.
ClusterOperatorDegraded started firing at 5:10:04Z.
ClusterOperatorDegraded stopped firing at 5:10:23Z.
The e2e suite complained about:
ClusterOperatorDownis similar, but I'll leave addressing it to a separate commit. ForClusterOperatorDegraded, the degraded condition should not be particularly urgent, so we should be find bumping it towarningand usingfor: 30mor something more relaxed than the current10m.