-
Bug
-
Resolution: Incomplete
-
Blocker
-
None
-
EKS 1.22
Controller 2.386 or 2.375.2 LTS
Helm chart inside EKS
I get very often websocket errors no matter what I do on shell steps even on latest inbount-agent with default settings. In order to improve my situation I have tried increasing the websocket timeout above 30 sec and it has no influence.
Helm values:
controller:
affinity:
nodeAffinity:
requiredDuringSchedulingIgnoredDuringExecution:
nodeSelectorTerms:
- matchExpressions:
- key: topology.kubernetes.io/zone
operator: In
values:
- eu-west-2a
javaOpts: >-
-Dorg.csanchez.jenkins.plugins.kubernetes.pipeline.ContainerExecDecorator.websocketConnectionTimeout=500
-Dorg.csanchez.jenkins.plugins.kubernetes.pipeline.websocketConnectionTimeout=360000
-Dkubernetes.websocket.timeout=280000
The overall websocket stays at 5 retries of a total 30 sec.
Jenkins pipeline is simple declarative call where sole stage step is to execute shell command if it is non shell this can't be reproduced.
The scope of the ticket is mainly aimed at the failure to modify the value for the timeout.