Repository navigation
A Service that keeps failing gets a new warning event every 30 seconds #45
Description
With #37, a Service that keeps failing gets a new warning event every 30 seconds.
robotlb publishes a SyncLoadBalancerFailed event on every failed reconcile. The event recorder in kube-runtime 0.96 creates a new Event object on each call and does not aggregate repeats, so a persistent error, for example a network that does not exist, leaves about 120 events an hour until they expire.
I'd like repeats of the same message to be skipped, or kube-runtime updated if a newer recorder aggregates them.
Activity
lexfrei commentedon Sep 29, 2026
kube 0.98 and later aggregate repeated events into a series when the same
Recorderis reused, so this mostly goes away with the kube update in #50. The chart role then needspatchon events too.