[Check_mk (english)] Calculating the derivative at point in a graph.

[Sebastián Sánchez]

Hello.

I'm monitoring a metric of a service.

There is some way to trigger an alarm by calculating the derivative
(dy/dx) at a point in the graph for some service. Using only check_mk
via wato.

Regards.

[Robert Sander]

Post by Sebastián Sánchez
There is some way to trigger an alarm by calculating the derivative
(dy/dx) at a point in the graph for some service. Using only check_mk
via wato.

No, not out of the box.

The check plugin that handles the agent data would need to compute the
derivative (there is a helper function get_rate() for that). Like the
if-check gets a counter of bytes, but the metric is in bytes per second.

If that metric is collected by an active check the Nagios plugin has to
compute the rate itself.

Regards
--
Robert Sander
Heinlein Support GmbH
Schwedter Str. 8/9b, 10119 Berlin

http://www.heinlein-support.de

Tel: 030 / 405051-43
Fax: 030 / 405051-19

Zwangsangaben lt. §35a GmbHG:
HRB 93818 B / Amtsgericht Berlin-Charlottenburg,
GeschÀftsfÌhrer: Peer Heinlein -- Sitz: Berlin

[Sebastián Sánchez]

Post by Robert Sander
If that metric is collected by an active check the Nagios plugin has to
compute the rate itself.

I'm doing it that way.
But I wanted to see if it could be done in another way.


Thanks.

[Steve Moulton via checkmk-en]

Sebastián,

I've had very good results from directing check_mk metrics into graphite
(there is a solid integration), and using grafana to display both atomic
and synthetic metrics (such as derivative, integral, etc) from the
graphite whisper (time metrics) database.

You can also export check_mk data into an influxdb database, which gives
you some greater capabilities.  I have found that maintaining influxdb
to be a larger effort though.   Whisper pretty much takes care of
itself, and is what I use in production.

I use check_mk interface for presenting administrative metrics, and
grafana for presenting performance metrics to users.

  - Steve

Post by Sebastián Sánchez

Post by Robert Sander
If that metric is collected by an active check the Nagios plugin has to
compute the rate itself.

I'm doing it that way.
But I wanted to see if it could be done in another way.
Thanks.

[Sebastián Sánchez]

Thank you Steve.