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}
;