Grafana Graphing

To enable Nagflux/Histou/Grafana performance graphs instead of PNP4nagios, execute the following steps:

omd config set PNP4NAGIOS off
omd config set GRAFANA on
omd config set INFLUXDB on
omd config set NAGFLUX on

When using icinga 2 as core, you have to enable the performance write by

icinga2 feature enable perfdata

Afterwards change the template imports from host-pnp or srv-pnp to host-perf or srv-perf for your configuration to use the correct action_url.
Or change the urls in the host-pnp and srv-pnp template, Whatever is easier in your case. For example:

define service {
# OLD
use srv-pnp,...
...
# NEW
use srv-perf,...
...
}

PNP4Nagios RRD Export

It is possible to import the existing RRD data into the influxdb by running the following script:

~/bin/migrate_pnp_to_nagflux

This might take a while, depending on the number of RRD files.

Dual graphing with PNP4Nagios and Grafana

It is possible to run both PNP and Grafana in parallel, for example during migrations or
for evaluation purposes. This is best achieved by using the Mod-Gearman module to duplicate
performance data.

If you use Mod-Gearman already, just add another perdata=grafana_data entry in the
neb module configuration etc/mod-gearman/server.cfg.

If you don’t use Mod-Gearman yet, switch it on using omd config and set services=no
and hosts=no in the neb module configuration etc/mod-gearman/server.cfg.

Then set pnp into gearman mode with:

omd config set PNP4NAGIOS gearman

and set Enabled = true in the ModGearman section of etc/nagflux/config.gcfg.
Also change the Queue to grafana_data.

Note you have to choose a primary graphing solution in thruk by using the right action_url in
your hosts / services. You can only use one here, either PNP or Grafana. However, you can
choose that freely for each host and service separatly.

Monitoring-Workshop 2017 12./13.9. Düsseldorf