One of my responsibilities in Red Hat is Management Console testing in Wildfly. In early development build of one of earlier releases we observed strange behavior in our test environment. Although the machine had a large amount of memory available at the beginning of tests, it always failed on java.lang.OutOfMemoryError when the test suite was running. We had to find out in which parts of tests this happens.

Since we are using Selenium and the tests are running in Firefox, the obvious solution would be to track amount of memory which the browser is using. But how? We agreed, that the ideal output for our use would be graph showing dependency of used memory on time.

After I googled a bit I found a question on stackoverflow.com posted by user who had similar problem. I though, that I could just use solution from the top voted answer or solution posted by Bruno Girin on his blog and that’s it. But it was not suitable for our usage. In our test suite there is used a brand new, clean instance of Firefox for each test. Because of that, I had to come up with solution which would track instances of Firefox which pops up on start of each test.

In Bash script, which can you find on my GitHub, output of ps command with -o option to format custom output is used to track these instances. Output in format <pid>,<memory-size>,<cpu-usage> is then saved in usage.log file by default. Path to target file can be changed by -l <PATH> option. Only other output is diagnostic log of script, which outputs to stdout and stderr by default so you can find out if there are any errors.

I documented all available options in readme and I believe that the code is simple enough to be self-explanatory if you need any further help. If you have not enabled creating graph simultaneously with log, you can create it anytime later with gnuplot -c gnuplot.script usage.log usage.png. In case you need bigger resolution, just edit first line in gnuplot.script and replace 800,600 in set term png small size 800,600 by desired resolution and run the command again.