This would help, but not fix #5120. It's now supports the running of multiple pods for scaling, without the issue that the session will not known on the backend pod.
@@ -62,7 +62,7 @@ objects:
targetPort: 8080
selector:
name: "${WEKAN_SERVICE_NAME}"
- sessionAffinity: None
+ sessionAffinity: ClientIP
type: ClusterIP
- apiVersion: v1
kind: Service