Given: A redis stream consumer application runs in Openshift/Kubernetes.
The service can be scaled and hence multiple consumers will run on different pods. To distinguish them a name should be given to each replica pod.
Is there a good practice or a working solution for that given the pods are pretty dynamic, their IPs are not static, their hashes change on each redeploy/restart.
How to correctly avoid orphan consumers preferably without any "zoo" keeper developed for this purpose unless there are tools for that? Is there a way to somehow stick to names or indices so at least scaling up would not leave consumer names in the group stuck?
