Publicly document what happens if ReadySet is down for a prolonged period #208
Description
When ReadySet is down for an extended period of time, the replication slot on the database will remain open and the WAL will consume more and more disk until ReadySet reconnects and catches up, which might take a while (this is the same behavior that would be exhibited by a normal read replica).
We should clearly document this behavior and explain that if ReadySet is down for an extended period, the deployment should be cleared out and the replication slot should be dropped. When ReadySet starts back up, it will create a new replication slot and resnapshot from scratch. The --cleanup flag can be used for this purpose.