I am having some trouble in making elastic scaling work with HA. I have a jobmanager pod and a single taskmanager pod, and I use a post install hook to start the jobmanager along with the job (As its required in application mode). If due to some reason my job manager goes down, and I start the same job from the required checkpoint, my HA (zookeeper) kicks in and submits another job to the jobmanager. Is HA not supposed to work with elastic scaling, or am I missing a crucial design piece here.