hystrix-to-librato

Sends Hystrix/Turbine stream metrics to Librato;

How it works

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).

Metrics sent

• hystrix.circuit.open: 1 if circuit open, 0 otherwise. Source will be {cluster}.{group};
• hystrix.latency.{lat}: The executition latency in ms. Source will be {cluster}.{group}.{name};