Issues with Autoscaling and Health check enabled in Openshift
We recently deployed one of the Pega applications (V8.2.4) into Openshift kubernetes as a containers. Everything went fine and go live was successful. However, users are experiencing the issues. Please find the details below.
Users are unable to transition over to the new pod in a seamless fashion, but they just got dropped off and had to re-login when we enable the options listed below.
we enabled auto scaling with the condition if CPU and MEM reaches above 80% for 3 concurrent times then Openshift automatically brings up new node and transition over to the new node from the old pod.
Openshift recommends to enable health check on pods. so we enabled in our PRODUCTION environment. The health check scenario is if pod reports unhealthy it should kill the pod and bring up the new node.
In the above both scenarios, users are getting dropped off from the existing pod and seeing home page to login again. But they are not able to move to the new pod without dropping from the current session like seamlessly. Please help in fixing the issue.
***Edited by Moderator Marissa to update platform capability tags; update SR details****
1. Can you double check if you have the following setting of value of 'true'? (check it from admin studio configuration screen)
2. In your deployment of web nodes, make sure you have at least 2 pods running at any given time. If you only have one pod and that pod is not healthy, there is no pod to move the saved session data to the healthy pod. The user re-login will be expected.
Yes Kevin. We do have the DSS setting enabled to true and yes we do have more than 3 pods for each environment. However, the issue is still exists. Are there any additional settings do we need to consider to apply to correct the behaviour ?...