Describe the solution you'd like

When I add a HTTP shovel that sends message to a http server, if for some reason that server doesn't respond or respond with a non-200 status I want to limit the number of retries, right now it will just keep trying until the endpoint is up again.

Some ideas are:

• exponential backoff: So it retries less often
• retry limits, drop the message if it has tried sending it for X times
• connect timeout: if the http server if down, limit the amount of time lavinmq should wait before rejecting the message

I guess this things should go as settings on the shovel.