Munin vs Ganglia vs Cacti vs Other

sdotsen

First thing first, how much RAM would I need to graph 5 servers? I'm probably going to throw Nagios on it as well so that I dont have to purchase multiple VPSes. I just want the basic, memory, cpu, network, etc...

I have experience with Munin and Ganglia but this was years ago and I had a 16GB host at my disposal :-). I'm looking for experience from others who have set up monitoring for their VPSes.

tchen

I'm running a zabbix server on a 256MB OVZ at about 120MB usage (with apache/mysql) and its monitoring about 24 servers. I think Nagios+Munin should be comparable, if not lighter since it uses just RRD.

sdotsen

If I recall Munin was a memory hog. Has this changed?

tchen

On another server a long time ago - I think it was a 128/256MB - I was running nagios+munin+nagiosgraph. I think I was monitoring 6 or 8 vms at that time fine.

If you're really concerned about sizing though, you could always look at librato metrics. That'd avoid most of the hassle.

vapornode

I have Munin setup monitoring over 30 servers right now. It runs on a 256MB LEB. The RAM usage stays at or around 35MB when idling and rises up to about 120MB when making the checks. The VPS it runs on also has 1 core which does go up to 100% when making the checks, but other than that, CPU usage is near 0%.

raindog308

How much RAM does munin take on the client?

herbyscrub

@raindog308 said: How much RAM does munin take on the client?

24030 0.0 0.9 7264 1296 ? Ss Feb02 0:10 /usr/sbin/munin-node

I run the graphing on a 64MB VPS against 10 servers with about 10 plugins each.

sdotsen

Thanks guys. Now I jut need to find an LEB in Dallas.

ryanarp

We use munin to show off our vps node performance in our signature. So far we are pleased with it.

Holoshed

I use Munin and Nagios monitoring about 10 machines on a 64meg vps:

USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
root 1 0.0 0.6 2028 432 ? Ss Jan22 0:30 init [2]
root 2 0.0 0.0 0 0 ? S Jan22 0:00 [kthreadd/1336]
root 3 0.0 0.0 0 0 ? S Jan22 0:00 [khelper/1336]
www-data 522 0.0 0.4 1900 284 ? Ss Jan22 0:00 /usr/sbin/fcgiwrap
root 550 0.0 0.2 5368 144 ? S Jan22 0:00 supervising syslog-ng
root 551 0.0 1.3 6008 896 ? Ss Jan22 0:47 /usr/sbin/syslog-ng -p /var/run/syslog-ng.pid
root 600 0.0 0.9 3808 596 ? Ss Jan22 0:17 /usr/sbin/cron
root 624 0.0 0.3 4992 224 ? Ss Jan22 0:00 nginx: master process /usr/sbin/nginx
www-data 625 0.0 0.5 5572 368 ? S Jan22 0:13 nginx: worker process
root 704 0.0 0.9 5772 644 ? Ss Jan22 0:36 /usr/lib/postfix/master
postfix 726 0.0 1.0 5924 664 ? S Jan22 0:16 qmgr -l -t fifo -u
root 772 0.0 0.6 2392 424 ? Ss Jan22 0:00 /usr/sbin/xinetd -pidfile /var/run/xinetd.pid -stayalive -inetd_compat -inetd_ipv6
root 863 0.0 2.5 7156 1692 ? Ss Jan22 3:48 /usr/sbin/munin-node
postfix 4608 0.0 1.2 5788 808 ? S Mar05 0:00 pickup -l -t fifo -u -c
nagios 7639 0.0 1.3 11036 900 ? Ssl Jan29 26:10 /opt/nagios/bin/nagios -d /etc/nagios/nagios.cfg
root 14438 0.1 1.6 8760 1076 ? Ss 00:06 0:00 sshd: root@pts/0
root 14442 0.0 2.2 4548 1476 pts/0 Ss 00:06 0:00 -bash
postfix 15187 0.0 0.9 6348 616 ? S Jan25 0:02 tlsmgr -l -t unix -u -c
root 16869 0.0 1.7 20388 1132 ? Ss 00:07 0:00 php-fpm: master process (/etc/php5/fpm/php-fpm.conf)
www-data 16871 0.0 1.0 20388 700 ? S 00:07 0:00 php-fpm: pool www
www-data 16872 0.0 1.0 20388 700 ? S 00:07 0:00 php-fpm: pool www
root 16970 0.0 1.0 5488 704 ? Ss 00:07 0:00 /usr/sbin/sshd
root 17135 0.0 1.5 3868 1020 pts/0 R+ 00:08 0:00 ps aux

free -m
total used free shared buffers cached
Mem: 64 51 12 0 0 32
-/+ buffers/cache: 19 44
Swap: 64 39 24

marrco

munin 2.0 with on demand graph and html generation.
see here: http://munin-monitoring.org/wiki/CgiHowto2

Raymii

I have munin 2 running on my raspberry pi monitoring about 25 servers. Not using on-demand graph generation (because, when I need to look at munin I don't want to wait, munin means problems with other servers.) but only once an hour.

https://raymii.org/s/tutorials/Munin_optimalization_on_Debian.html

This guide also covers rrdcache, ionice and tmpfs.