shrine

Pull request #1529: Feature/SHRINE2020 1218 cats effect upgrade

Merge in SHRINE/shrine from feature/SHRINE2020-1218-cats-effect-upgrade to develop

* commit 'ad866f30bbb28a9192c4de8892d2d97ac70ce51f': (92 commits)
  cats effect IO's theads are daemon by default - no reason to do anything to get that - SHRINE2020-1655
  Added back the LocalMessageQueueStopper stop call for SHRINE2020-1653
  Restoring the breakdowns to the jmeter file for SHRINE2020-1654
  Refactored the legacy client to handle both PM cell and CRC cell requests for SHRINE2020-1654
  Small profound changes to fix problems at scale for SHRINE2020-1654
  Removed soem uninteresting logging for SHRINE2020-1654
  Try a separate semaphore limit for PM cell calls for SHRINE2020-1654
  Moved the timeout to the right spot and got rid of the unneeded execution context for SHRINE2020-1654
  Use a semaphore to limit the number of i2b2 calls (instead of a thread pool) for SHRINE2020-1654
  Removed the unused implicit from RetryIO for SHRINE2020-1652
  Tweak RetryIO in a chaser ticket
  Cleanup for SHRINE2020-1650
  Got messages logner than 4k sending for SHRINE2020-1650
  More logging to maybe bracket the problem for SHRINE2020-1650
  More logging to maybe bracket the problem for SHRINE2020-1650
  Some more logging to examine SHRINE2020-1650
  Code review suggestion from Xav for SHRINE2020-1642
  i2b2 really doesn't behave well when it is overloaded. I added back shrine's internal timeout for SHRINE2020-1637
  Moved the timeout for connections to the actual http client, and put the gateway execution context in place for SHRINE2020-1637
  Renamed ShrineServletContextSyntax to ShrineServletMounter and cleaned up some associated code for SHRINE-2866
  ...