You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Support for dynamically defining the ProbeFields on a per placement or managed cluster basis.
Use case
The observability group continues to develop the MultiClusterObservabilityAddon (MCOA), this addon has the ability to selectively deploy different observability signals collectors to different spoke clusters. Some observability deployment can be rather complex so we would want the ability to let users know if their deployments have been successful. To achieve this we started using the AgentHealthProber of type HealthProberTypeWork. Unfortunately this probe only supports a fixed list of ProbeFields which is not sufficient for our use case given some clusters might just want metrics while others might want metrics and logs.
The text was updated successfully, but these errors were encountered:
This requirement is similar with #287
I have a PR #289 to address this.
In the PR, support wildcard in the name and namespace of ProbeField.ResourceIdentifier (depend on this open-cluster-management-io/ocm#705), and add another func HealthChecker to WorkHealthProber which can check all feedback results at once. @JoaoBraveCoding could you help check if this can address your case?
@zhiweiyin318 thank you for the quick work! I've already drafted a PR and tested this and it seems to do the job! I'll double check with the team to be sure 👍 Great work! 🙌
Request/Feature
Support for dynamically defining the
ProbeFields
on a per placement or managed cluster basis.Use case
The observability group continues to develop the MultiClusterObservabilityAddon (MCOA), this addon has the ability to selectively deploy different observability signals collectors to different spoke clusters. Some observability deployment can be rather complex so we would want the ability to let users know if their deployments have been successful. To achieve this we started using the AgentHealthProber of type
HealthProberTypeWork
. Unfortunately this probe only supports a fixed list ofProbeFields
which is not sufficient for our use case given some clusters might just want metrics while others might want metrics and logs.The text was updated successfully, but these errors were encountered: