As part of STREAMS-1141, we want to be able to rely on expiring checkpoints from object storage based on TTL settings. Once we start doing that, in theory, it is possible that a long running but idle processor takes no checkpoints and so we end up in a situation where we may not have a checkpoint for it in object storage.

Tweak the checkpoint coordinator so that we force a checkpoint after 7 days since the last checkpoint even if nothing has changed.