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
Found an issue with kafka nodes are coming up again after having failed. If they wake up to a corrupted index, it will attempt to fix itself. This seems to have two major implications:
Memory consumptions goes out the roof, resulting in pod getting killed due to OOM (due to limits, of course)
Readiness probe fails, and will kill the pod if it hasn't already OOM'ed