Warn about liveness probe for Zulip migrations
Xeruf performed a migration from standalone Zulip to Stackspin-Zulip. We should add this to the documentation:
Downtime is not the problem, the issue is the following: After running ./manage.py delete_realm -r '' the health checks for the pod will report it as broken, so if you don't run the import fast enough, the pod will be killed, a new one created, and you have to start all over.