stop insert retry and drop queue after too many retries

[oryband]

I experienced a scenario where my elasticsearch cluster had a downtime. This caused my elastigo client to retry inserts over and over again. It was queuing more and more documents for sending while it was sleeping between retries. Eventually the RAM filled up and the machine stopped responding until the app running elastigo was restarted.

Please add a feature for max retries, and let elastigo drop it's current insert attempt and move on.

[oryband]

this basically comes down to providing some sort of circuit breaking, should probably be done with exponential backoff as well