Add retry mechanism for known exceptions

[Rictus]

Hello @codemation ,

In easyauth, a gunicorn server is launched and some other code here in easyrpc "wait" for it. But in case of failing (one of KNOWN_EXCEPTIONS) I do not see a retry mechanism. Maybe a miss?

Maybe the retry mechanism should be in easyauth but as I've seen the comment # retry in easyrpc, maybe that's the correct location.

Tell me what you think. If this is accepted, we could remove the asyncio.sleep(5) at line 254 of easyrpc/server.py which feels random and is not enough wait time in my setup.

Thanks!

[Rictus]

Did you had any chance to take a look to it @codemation ?