Add EventRecorder to Helper #320
Conversation
|
I don't have much context about events in k8s. I think it worth to create a doc where you propose what will be the use of the events in openstack-k8s-operators. I.e. what step during a reconciliation worth emitting an event. What information an event should always contain, etc. A bit similar to the definition of the condition usage https://github.com/openstack-k8s-operators/docs/blob/main/conditions.md |
I agree with @gibizer , I think we should not just create events, first we'd have to agree/define when there should be one and the content. |
|
I think this can help us to detect race conditions, we can sort events and see the chronological order things were added/updated in the cluster |
fao89
force-pushed
the
event
branch
2 times, most recently
from
0b59b50 to
b23188e
Compare
fao89
force-pushed
the
event
branch
2 times, most recently
from
5c1988f to
5271d47
Compare
Depends-On: openstack-k8s-operators/lib-common#320 Signed-off-by: Fabricio Aguiar <[email protected]>
Depends-On: openstack-k8s-operators/lib-common#320 Signed-off-by: Fabricio Aguiar <[email protected]>
Depends-On: openstack-k8s-operators/lib-common#320 Signed-off-by: Fabricio Aguiar <[email protected]>
Depends-On: openstack-k8s-operators/lib-common#320 Signed-off-by: Fabricio Aguiar <[email protected]>
fao89
force-pushed
the
event
branch
6 times, most recently
from
c7345be to
ab9c8a7
Compare
Depends-On: openstack-k8s-operators/lib-common#320 Signed-off-by: Fabricio Aguiar <[email protected]>
https://www.cncf.io/blog/2023/03/13/how-to-use-kubernetes-events-for-effective-alerting-and-monitoring/ https://www.cncf.io/blog/2021/12/21/extracting-value-from-the-kubernetes-events-feed/ https://www.atlassian.com/engineering/how-to-export-kubernetes-events-for-observability-and-alerting Signed-off-by: Fabricio Aguiar <[email protected]>
Depends-On: openstack-k8s-operators/lib-common#320 Signed-off-by: Fabricio Aguiar <[email protected]>
Depends-On: openstack-k8s-operators/lib-common#320 Signed-off-by: Fabricio Aguiar <[email protected]>
Depends-On: openstack-k8s-operators/lib-common#320 Signed-off-by: Fabricio Aguiar <[email protected]>
| @@ -105,8 +105,12 @@ func createOrPatchConfigMap( | |||
| return nil | |||
| }) | |||
| if err != nil { | |||
| h.GetRecorder().Event(obj, corev1.EventTypeWarning, "ConfigMapError", fmt.Sprintf("error create/updating configmap: %s", cm.Name)) | |||
There was a problem hiding this comment.
re ConfigMapError and other events, its probably good that we create a central reason collection which then gets used in all the operators. similar to how we do it with the conditions. Otherwise I guess if events are used in the service operators, we'll see different reasons used. I could see two ways. Either we create an events pkg and have the consts for reasons in there, or they are part of the already existing pkgs, like ConfigMap reason in configmap. but then how to handle reasons for stuff we do not yet have. create extra pkgs, or better the one events pkg. thoughts?
| return ctrl.Result{}, err | ||
| } | ||
| if op == controllerutil.OperationResultCreated { | ||
| h.GetRecorder().Event(h.GetBeforeObject(), corev1.EventTypeNormal, "DaemonSetCreated", fmt.Sprintf("daemonset %s created", d.daemonset.Name)) | ||
| } | ||
| if op != controllerutil.OperationResultNone { | ||
| util.LogForObject(h, fmt.Sprintf("DaemonSet: %s", op), daemonset) |
There was a problem hiding this comment.
should there be an updated event when the object got updated?
I'm worried that this PR is moving forward without agreeing on the scope of it. |
agreed, @fao89 do you plan to work on some definition doc as we have for conditions? |
|
yes, I started it here: openstack-k8s-operators/dev-docs#105 |
|
Actually, after sorting the events on must-gather(openstack-k8s-operators/openstack-must-gather#47), I can see the events I was interested in are already there (pod/job creation and job completion): |
|
by the community guide: I think we can close this PR |
https://www.cncf.io/blog/2023/03/13/how-to-use-kubernetes-events-for-effective-alerting-and-monitoring/
https://www.cncf.io/blog/2021/12/21/extracting-value-from-the-kubernetes-events-feed/
https://www.atlassian.com/engineering/how-to-export-kubernetes-events-for-observability-and-alerting
Events: https://gist.githubusercontent.com/fao89/f25d02317adfebee8943055be23a4553/raw/cf40e9197b6507a67722f334722ac5fa6c382270/events.log