Issue
We are running into situations where "write" API calls begin to hang indefinitely/fail.
So far we have noticed this on the functions create sidecar and cancel sidecar functions. However, things like quoting orders, and viewing previous batches still works.

I cannot determine whether the issue occurs, and a user hits it, and then we notice the issue,

OR

the user creates an order, and something bad happens during their registration process, and in the process, it breaks our Pool instance.

Expected behavior

The user should be able to register a sidecar, and if something goes wrong, things should cleanup smoothly, and not break our Poold instance.

**Potentially related error logs: **
This error was being put into our logs, for a variety of different sidecar tickets, relatively frequently, around every few seconds.

2022-01-20 19:24:01.276 [ERR] AUCT: [/poolrpc.HashMail/RecvStream]: rpc error: code = Unknown desc = stream not found
2022-01-20 19:24:01.276 [ERR] POOL: unable to recv cipher box msg: rpc error: code = Unknown desc = stream not found

After restarting our poold instance the above errors were not spammed in our logs anymore.

However I noticed this message in our logs, after restart

2022-01-20 20:57:52.345 [ERR] POOL: unable to transition state: unhandled provider state transition ticket=10c836837ba5021a, state=offered
2022-01-20 20:57:52.354 [ERR] POOL: unable to transition state: unhandled provider state transition ticket=1108335ff3420a29, state=offered
2022-01-20 20:57:52.376 [ERR] POOL: unable to transition state: unhandled provider state transition ticket=2b6b384a17f3f213, state=offered
2022-01-20 20:57:52.384 [ERR] POOL: unable to transition state: unhandled provider state transition ticket=310358470b8866f1, state=offered