Gradually decrease "connecting..." log message frequency on peers

Peer currently logs pretty frequently when it doesn't get a websocket connection to master. This can fill up RAM/log with unnecessary information. Fix: gradually decrease frequency it writes that log message (to min once per 30min) and then reset the timer when something else happens.

Start by locating the peer websocket connection code and the "connecting..." log message. Trace how connection failures and other peer events are handled; done means repeated failures are logged no more often than once per 30 minutes, while another event resets the timer.

Written by the indexing model from the issue text.