Improve detection of when reddit api service is down and take better action

[FoxxMD]

Recent reddit api issues have caused abnormal or unexpected behavior with CM. Need to make a concerted effort to detect "bad api state" better and fallback to stopped bot state faster and more gracefully.

• [x] Turn off or modify continueAfterRatelimitError behavior in snoowrap
  • Need to catch rate limit error at bot level and put bot in stop state while waiting for limit to reset
• [x] Review retry handler for polling to ensure it isn't unexpectedly looping or not exhausting retry attempts somehow
• [x] Better handling of api errors during cold start
  • Bots/server should still be able to be initiated and started up even if initial api testing fails. We should be able to handle re-creating api client with ui action on bot restart

[FoxxMD]

Addressed most of these in b2846efd2ba4e8fd76ddcea8ee37a9c6cb8f5c3b and 4b14e581dd140c617588fbee4a6df741bbfd8102 :

• Turn off snoowrap request queuing
• Aggregate errors from all managers at bot-level and force all to stop (and clear queue/polling) if a small threshold is met
• Add activity refresh on check into try-catch so delayed activities in queue don't cause a loop if they fail due to api issues
• Sanity check api call to /api/v1/me on heartbeat to see if anything accessible before trying to restart any managers
• Stagger manager restart on heartbeat

See the linked issue below for a roadmap on further improvements

[FoxxMD]

In 5e70ca1cb6423ebb12b2e05a55c6b858bf21d03b :

• Fix polling timeout to actually stop on error by simplifying timeout and waiting until response is OK to recreate next timeout call
• Use "unexpected exception" retry count for all non well-known "reddit blip" responses in retry handler rather than failing immediately AND log this distinction
• Fix managers not emitting errors from checks
• Fix bot not awaiting retry handler on manager error emit
• Increase nanny loop delay on error to reduce api pressure when there are many bots running