Status codes from GES do not always indicate what they are interpreted as

[litch]

My cluster entered a state where it would reject all writes that wound up getting routed to a slave via DNS (which was "fun") and they return 400 Bad Request, which is then interpreted as WrongExpectedEventNumber - probably because the documentation is incomplete.

I described it here: https://groups.google.com/forum/#!topic/event-store/LE473sLEk_c

There is also a case pretty often where I get a 307 in response to a write while my cluster is in an error state. No clue why.

So I think we need to do some more parsing of the responses, and can't rely on status codes. Sadly.

[rybex]

Thanks for your last PRs ;) I will take a look on this issue and create ticket in our TODO list ;)