We have two clusters with multiple namespaces in each project. However our Error Reporting groups all errors together in the whole project and marking them as ‘gke_instances’ - there must be a better way to track where is the error coming from exactly (cluster/namespace/workload name)?

  • Thank you for posting a question. Could you modify the question so that it contains the attempts you have done in order to solve the issue?
    – 030
    Dec 24 '19 at 13:33
  • The 13 Dec 2019 error reporting release notes suggest it may now replace "gke_instances" with pod labels k8s-pod/serving_knative_dev/service and k8s-pod/serving_knative_dev/revision. I'm yet to try, but I stumbled on this question while looking.
    – icio
    May 26 '20 at 9:31