We are currently running Nagios to monitor our server performance, however, recently we are trying to evalute two chat modules (one from the 1.6 beta, the other AJAX-based) with their scalability and performance. As Nagios is kind of difficult to set up (probably same with JFFNMS and OpenNMS), I wonder if there is any other light-weighted program (maybe SNMP based) that we could setup to monitor the server health or CPU/memory load when we setup automated robots (bots) to hit the two chat modules?? Any suggestions or comments? Many thanks!
I am not sure how I should call you "Wen" or "Wen Hao" or "Chuang". Sorry, it's difficult for me to identify the surname and the first name.
I do not have an answer to your question but I would like to know which AJAX chat you are testing.
Thanks for your answer
We are testing the lace, which could be found here:
We are in the process to integrate it with Moodle (we are running 1.5.2+ here at SFSU), would like to test this later with the improved chat module that came with 1.6 beta...
Sorry to reply late..
What we normally do is we write some stats out from admin/cron.php to a stats file in moodleroot. For example: how many unique users have log entries in mod_log in the last 5m. For chat profiling, you could record how many chat messages and unique users in the last 5m. It is much easier in Postgres and recent versions of MySQL because you can use subselects and just get the COUNT
I've been using Nagios for the past 3 years but I'm moving to JFFNMS. I think JFFNMS is easier to setup than Nagios and if you run it on Suse Linux you can follow the installation instructions and it will work right away. I'm actually running JFFNMS on a vmware virtual computer and have plans to release it as a virtual appliance. The best thing to do is to follow Martin Langhoffs advice about the script to get stats from admin/cron.php and to correlate the result with the stats from JFFNMS.