Which monitoring system would people recommend?
by Lockywolf from LinuxQuestions.org on (#5P64G)
I have been using SmokePing so far, which "kinda" worked, and let me know whether my machines are dead or alive.
Althought it is not without its benefits, recently I started to find it inconvenient:
1. It really consumes a lot of RAM. Basically, it launches a Perl process per probe, and I have quite a lot of probes, so having ~1G of RAM used by monitoring is not nice at all.
2. It is only monitoring network. Although Smokeping can be tricked into running other programs at specified intervals, it is not really fit for presenting anything but network latency. I would like to see cpu usage, memory consumption, perhaps some SMART stats, like disk temperature.
3. It's quite fiddly. I had to really put in some crutches to, e.g., restart stuck processes.
SlackBuilds.org has quite a lot of scripts for monitoring systems, and in fact, there are too many. Right from the top of my head I can remember Zabbix, Nagios, Icinga, Cacti, Munin, Prometheus, and, to a lesser degree, atop, ntopng, nvstat, darkstat.
It would be hard to try all of them, so I would appreciate Slackware community opinion on which one is better.
I really liked the ability of ntopng to analyze network flow and similar things, but it comes with a prices -- I cannot spend 500Mb of RAM on each of my machines, and their nProbe is closed-source and not cheap.
Althought it is not without its benefits, recently I started to find it inconvenient:
1. It really consumes a lot of RAM. Basically, it launches a Perl process per probe, and I have quite a lot of probes, so having ~1G of RAM used by monitoring is not nice at all.
2. It is only monitoring network. Although Smokeping can be tricked into running other programs at specified intervals, it is not really fit for presenting anything but network latency. I would like to see cpu usage, memory consumption, perhaps some SMART stats, like disk temperature.
3. It's quite fiddly. I had to really put in some crutches to, e.g., restart stuck processes.
SlackBuilds.org has quite a lot of scripts for monitoring systems, and in fact, there are too many. Right from the top of my head I can remember Zabbix, Nagios, Icinga, Cacti, Munin, Prometheus, and, to a lesser degree, atop, ntopng, nvstat, darkstat.
It would be hard to try all of them, so I would appreciate Slackware community opinion on which one is better.
I really liked the ability of ntopng to analyze network flow and similar things, but it comes with a prices -- I cannot spend 500Mb of RAM on each of my machines, and their nProbe is closed-source and not cheap.