This was a problem with the JNLP service in kubernetes. As a node port service, it was doing round-robin through the available IPs, but would only connect if it happened to get the IP of the worker node running Jenkins. Any other IP would cause it to time out and retry. The problem would get worse the more worker nodes were present in the cluster.
I solved the problem by changing the service to a headless clusterIP, which is possible only because EKS gives pods VPC-native IPs. I'm not sure how I would solve this problem in a Kubernetes cluster using some other CNI that gives the pods addresses outside of the VPC range.
This message was sent by Atlassian Jira (v7.11.2#711002-sha1:fdc329d)
You received this message because you are subscribed to the Google Groups "Jenkins Issues" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/d/optout.