Bacula per job statistics
About
The bacula_job plugin allows the munin monitor to graph statistics of the Bacula backup system. The bacula_job plugin specifically charts the read throughput of the configured bacula-fd clients.
The plugin works by calling the bconsole program and screen scraping the output.
Screenshot
Download
Installation
- Make sure the bconsole binary is accessible in your path and the status command works as expected.
- Copy or link the bacula_job script into your munin-node plugin directory, usually /etc/munin/plugins.
- Configure the munin-node to run the bacula_job plugin as user root or another user able to execute bconsole and connect to the director:
[bacula_*] user root
- If you want the munin plugin to output the hostname of the director, set the munin environment variable report_hostname to yes:
[bacula_*] user root env.report_hostname yes
- Test the plugin by calling munin-run bacula_job config and by executing munin-run bacula_job. Both commands should output sensible data without failing.
Notes
- The plugin is best run on the same system as the director. This way, no network latency is introduced into the measurement as the connection is considered to be local.
- Due to the nature of bconsole, only jobs currently running while the munin plugin is executed are counted. This means short jobs which are started and finished between two munin runs are not tracked at all.
- Due to the same design limitation the reported read at the beginning and the end of a backup job might be a bit off.
Author
The munin plugin was written by Andreas Thienemann for bawue.net.
License
The code is available as free software under the GPL Version 2.
Bugs and Contact
In case of bugs, please either open a ticket in this Trac instance or write to <andreas (at) bawue (dot) net>.
