Performance Data Inaccurate for "check_xi_service_http"

[phaladi]

Good Day.

We are monitoring some ports and an FQDN using "check_xi_service_http".

Please see attached screenshot. This shows the performance graph view I am referring to.

When the FQDN or a port does not respond the green line disappears for the duration of the downtime though returns once the issue has been resolved.

This does not give us any indication of when and for how long the port or FQDN was down as the green line is filled in for the time the service was down.

Is there a way to record the downtime in the performance graph?

Kind Regards,
Gary.

[cdienger]

I just tested this out and not aware of any customers doing this so if you do take these steps I would recommend doing so on a lab machine first.

/usr/local/nagios/etc/pnp/process_perfdata.cfg contains a line:

RRD_HEARTBEAT = 8640

If a service is down longer than the heartbeat value then it will log NaN for these data points once the service comes back up. Lowering the heartbeat:

RRD_HEARTBEAT = 120

and deleting the current database files(/usr/local/nagios/share/perfdata/<hostname>/<servicedesc>.[rrd|xml]) to allow a new database to be created with this setting should produce more desirable behavior.