-
Bug
-
Resolution: Unresolved
-
Undefined
-
1.37.0, 1.38.0
-
None
-
None
-
False
-
-
False
-
-
When new ksvcs are modified (or added/removed), with larger number of ksvcs on the cluster (~500ksvcs),
the LDS updates in 3scale-kourier-gateway take longer than 1s, the gateway fail readiness/liveness probes (and possible also blocks serving traffic), causing restarts of 3scale-kourier-gateway due to liveness failures.
With internal-tls-encryption enabled, during a script creating 100 namespaces with 10 ksvcs each (10ksvcs at a time, then waiting for their readiness)
Events: Type Reason Age From Message ---- ------ ---- ---- ------- Normal Scheduled 13m default-scheduler Successfully assigned knative-serving-ingress/3scale-kourier-gateway-6cb499dc6d-s8cl9 to maschmid-1178-xwr8p-worker-0-j7mjj Normal AddedInterface 13m multus Add eth0 [10.129.5.142/23] from ovn-kubernetes Warning Unhealthy 11m (x11 over 13m) kubelet Liveness probe failed: Get "http://10.129.5.142:8081/ready": context deadline exceeded (Client.Timeout exceeded while awaiting headers) Warning Unhealthy 3m32s (x45 over 13m) kubelet Readiness probe failed: Get "http://10.129.5.142:8081/ready": context deadline exceeded (Client.Timeout exceeded while awaiting headers) Normal Killing 113s kubelet Container kourier-gateway failed liveness probe, will be restarted Normal Pulled 96s (x2 over 13m) kubelet Container image "registry.redhat.io/openshift-service-mesh/proxyv2-rhel9@sha256:fe1e78971352ddd76ee6f88aa0f4e4ac6d1298da8c960f877ae9f883c291dfd6" already present on machine Normal Created 96s (x2 over 13m) kubelet Created container: kourier-gateway Normal Started 96s (x2 over 13m) kubelet Started container kourier-gateway
