Sends Hystrix/Turbine stream metrics to Librato;
You pass down all configs via environment variables and start up the binary.
It will create a goroutine for each cluster being monitored, and will log each report and how many goroutines are alive.
If any HTTP error occurs, it will try to restart itself in 5 seconds.
The reports to Librato are also made in new goroutines, but the same metric will be sent at most every 5 seconds (to avoid paying too much, since librato charges by metric sent).
hystrix.circuit.open:1if circuit open,0otherwise. Source will be{cluster}.{group};hystrix.latency.{lat}: The executition latency in ms. Source will be{cluster}.{group}.{name};