Please fill in the following information:

Issue description: The deployments that are dropped during this labs setup have a good chance of practically killing one of the nodes. It would be a good idea to dial down the amount of load dropped on it or changing what kind of alert is supposed to appear.  

I’ve experienced this twice.
Worker03 simply wasn’t able to deal with the load to a point where cluster operators on this node would not respond and node was marked as unavailable. Given time maybe it would recover, but I gave it 30m to no avail. In the end I force deleted the pods and restarted the node. The problem is that because the cluster lost the communication with the node it didn’t even produce the desired alert

Steps to reproduce:

Workaround:

Expected result: