All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
[TUTORIAL] - Collectd client, server, web frontend tutorial, including Ansible playbook
What is Collectd?
What is collectd? collectd gathers statistics about the system it is running on and stores this information. Those statistics can then be used to find current performance bottlenecks (i.e. performance analysis) and predict future system load (i.e. capacity planning). Or if you just want pretty graphs of your private server and are fed up with some homegrown solution you're at the right place, too ;).
Demo
http://vps5.sparklingclouds.nl/collectd/ and http://vps21.sparklingclouds.nl/collectd/
Server tutorial, includes the nice web frontend
https://raymii.org/s/tutorials/Collectd_server_setup_tutorial_with_web_frontend.html
Client tutorial
https://raymii.org/s/tutorials/Collectd_client_setup_tutorial.html
Ansible playbook (both server and client)
https://github.com/RaymiiOrg/ansible
Collectd homepage
LEB Exclusive commentry
I'm looking for a complementary software to monitor my servers, now only munin and nagios, and I thought Collectd would fit nicely in the picture. However, I couldn't fine any tutorials covering an easy setup for both server, client and web frontend, so I wrote one myself.
LEB style, it uses only one thread, and the server part that runs on vps5 (the demo) is a BuyVM 128MB. You can see how it is performing fromout collectd graphs itself. The nice part is that this has on-demand graphing working right away, not like munin. It however does not has a nice overview like munin, but that I can live with.
The clients are all ranging from 32MB to 256MB ram max, and performing fine. It does however pull in an awfull lot of dependencies...
Go and try collectd, see how it works, and let me know what you guys (and gals) think about the tutorial and Collectd?
Comments
Looks good.
re: "pull in an awful lot of dependencies". Going a level deeper, probably worth explaining that Debian breaks out non-core collectd plugins into separate packages and by default installs rrd related packages, even on clients (!).
A #lowend approach:
apt-get --no-install-recommends collectd
This will install collectd and collectd-core, but skip all the other non-essential collectd plugins and underlying libraries. This is great for collecting/sending core statistics (cpu/ram/disk etc) and maybe sufficient for a LEB use case. If someone needs to monitor a MYSQL instance, they can apt-get libmysqlclient... How do you know which package to pull in? Look at the "recommends" section of "apt-get show collectd" and find the package name containing the relevant term...
collectd is awesome and kudos for writing up a guide!
Why not just use something like Observium?
While I'm a bit new to SNMP monitoring solutions, so feel free to correct me, SNMP appears quite limited to me. Good information but difficult to customize and break down further.
Good tutorial @Raymii. You should write on something on Nagios sometime
You can run collectd and unix-agent data through Observium quite easily The amount of information you can pull out is pretty amazing.
Just looked at the ansible playbooks you created for this - very cool. If anyone is getting into ansible for config mgmt, these are great examples to learn from.
@jarland
SNMP is basically limited by the MIB's if such, and as @MartinD stated Observium can run with somehing like check_mx.
Lately I'm also checking out Zabbix, and it looks good. Basicly combined Nagios & Munin plus a GUI to configure it all... Might be a bit too much however...