- Verify that the user application continues to work without errors during a Pod crash.
- Verify that new Revisions can be created for the user application.
- Verify that the user application can be scaled using concurrent requests metric.
- Verify that the user application can be scaled from zero.
- Verify that
SRVKS-385does not happen. In other words, killing a pod should not lead to stopping all application pods and starting them again with a new scale request.
As opposed to other scenarios where we kill the pod and then verify all functionality of the autoscaler, try to kill the pod during the scale up/down process (scale down but not to zero) and check there are no failed requests.
- is related to
-
SRVKS-700 Implement Knative autoscaler HA configuration in operator
-
- Closed
-
