We should provide a plugin for Nagios (http://www.nagios.org) to monitor
ThinLinc cluster operations. This would probably mean some extensions of
VSMserver and/or VSMagent as well. There should also be a document with
recommendations for how to monitor ThinLinc.
Things that are interesting to monitor include:
* SSH responding on all VSM servers and -agents. This is possible with existing
plugins, so it should just be mentioned in a document or HOWTO.
* License status. How many licences are available and used?
* Cluster connection status. Can the VSM server connect to all it's agents.
* Cluster load. Are all agent machines running at full load all the time,
meaning more machines should be added?
* VSM agent machines should respond on their port and be able to tell their load
* The document should describe how to monitor available disk space and load on
the VSM server machine(s).
* The document should describe how to find out if a HA setup has failed over to
the secondary node.
One way to implement this bug is to add support to the real tlclient for running in some kind of headless/scripted mode.
The earlier check_thinlinc_session has now been rewritten to use the "new" VSM API.
For simple up/down check on vsmserver, we can reuse the Nagios check_http plugin:
> /usr/lib/nagios/plugins/check_http -H <hostname> -p 9000 -u /RPC2 -j POST -P "<?xml version=\"1.0\"?><methodCall><methodName>fake</methodName><params /></methodCall>" -T text/xml -s 'Go away'!
Replace -p 9000 with -p 904 for checking an agent.