Currently when an instance is marked to be deleted, is tagged with value=timestamp.
We might want to wait a certain time for the node to be drained, but depending on the situation this might never happen (or might take a really long time).
For that scenario, we want to wait a certain time for the instance to be drained. If after this time the instance still has tasks from protected frameworks, we should kill the instance anyway.
Currently when an instance is marked to be deleted, is tagged with value=timestamp.
We might want to wait a certain time for the node to be drained, but depending on the situation this might never happen (or might take a really long time).
For that scenario, we want to wait a certain time for the instance to be drained. If after this time the instance still has tasks from protected frameworks, we should kill the instance anyway.