I work for a company that is looking to replace it's server monitoring system with an open source tool. They only have Linux servers to monitor. So far, so good.

The "killer feature" of the existing tool, from management's point of view, is the ability to generate reports (think a PDF document with nice colourful charts and tables) giving a monthly overview of various statistics collected by the monitoring tool. This extends from the obvious (uptime, network link usage capacity...) to the more subtle indicators, that inform about the "IT system's overall health" (ie, number of times the backup script failed, % of time when a non-production service was down...). This last section is important to them: information about less visible faults, that can provide them with information as to "how IT is in general", rather than just "are the services up?".

I'm looking for any opinions or suggestions on what open source tools out there you think would be good to set such a reporting system up. Open source is mandatory, and being able to configure the reports so that they are automatically generated is too.

Thanks for any hints!

We use Icinga here at work. Icinga is a fork of Nagios and 100% compatible with all add-ons for Nagios 3.

https://www.icinga.org/

It has integration with Jasper for reports though we don't bother with that feature and we avoid the PHP front end. As far as historical statistics go, at home I use PNP4Nagios with Icinga. At work we use Munin to track historical stuff. PNP4Nagios has a PDF export feature which is another report.

As far as "number of times the backup script failed" you'll have to write your own plugin for that since you're the only one who knows how your backup system works. You're going to find that's the case with any monitoring system.

Other open source reporting check out the thread for Network Monitoring of the year. There's a large list of open source monitoring solutions there.

Thanks for the pointers.

I should probably clarify one point: I'm well aware I need to write my own plugins for backups, my point was rather that I need reporting not on the associated *time* that is considered by the tool as "not OK" ("down time"), but the *number* of times it has entered that state. As my backups run daily, the "down time" would always be about 24 hours, which is uninteresting, however knowing that there were 14 incidents when backups failed over a month can show a trend.

Hope I'm making myself a bit clearer