Graphing of checks / nodes

[Nico Schottelius]

Good morning,

I have seen the telemetry support in consul for streaming data about
consul itself.

I was wondering if anyone (I guess so) is actually streaming / graphing
the results of checks?

My use case are the following:

- have consul check the diskspace/memory/cpu usage
- forward the metrics (statsd?)
- create graphs
- list number of nodes per datacenter

My current idea was to take something like statsite + grafana for this.

Any pointer on how you have done / would do this is welcome!

Cheers,

Nico

--
New PGP key: 659B 0D91 E86E 7E24 FD15 69D0 C729 21A1 293F 2D24

[Michael Fischer]

You could make checks return metrics in their standard output à la Nagios (see http://nagios.sourceforge.net/docs/3_0/perfdata.html), which Consul would expose in the service status.  At that point you could read the health API to collect the output, and process as you please.

Alternatively, if Consul is configured to run health checks at an interval, you could make the service check scripts gather metrics and write them into the node's statsd instance as a side effect.  This would be a simpler, more elegant approach IMO.

Best regards,

--Michael

--
You received this message because you are subscribed to the Google Groups "Consul" group.
To unsubscribe from this group and stop receiving emails from it, send an email to consul-tool...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.