All performance testing tools that I know of are bad at collecting system metrics from Unix-like system (AIX, Solaris, Linux, BSD etc). If you are lucky you might (might!) be able to collect CPU utilisation and maybe one or two memory utilisation counters.
Generally you have to create a shell script to run perfstat (or similar) then, when your load test is over, you process the output file and import it into your analysis tool.
Yesterday I learnt that you can write whatever data you want to an SNMP MIB so, if your performance test tool is able to gather metrics from SNMP, it follows that you can write your performance stats to the SNMP MIB and cut out the painful process of processing and importing for each test.
A little extra research shows that some operating system versions already have MIB objects with performance counters for metrics like CPU and memory use, although collection of these may have to be enabled. The OpenSTA performance testing tool website even has an article on “Monitoring CPU usage with OpenSTA via SNMP“.
This is a useful technique to add to any performance tester’s repertoire.
Comments are closed.
Check out the Slashdot thread…
Ask Slashdot: Performance Monitoring for Linux
muadib wants to know about the following: “Given the current discussions on tuning, I am trying to find out if there are any performance monitoring applications for Linux. I don’t mean things like xload, xosview, etc which provide only a small amount of data. For anyone who’s done benchmarking under NT, I mean something like their built in perfmon utility that lets you view and capture just about any statistic on your system or on a remote system. Capturing is the specific functionality I’m looking for b/c I’m working on a Linux device driver, and it would be nice to have historical data of CPU utilization, interrupts/s, etc. so that I can compare complete system perfomance between code revisions.”
For performance monitoring with SNMP on Windows check out this support answer for WhatsUp Gold.
I’ve popped a note on my blog linking to an article I wrote on UNIX monitoring during load tests. I agree with the feelings expressed above – sounds like we’ve been over similar ground!
If you’re interested check out: