stress/cpu_stress_ng_test stuck on specific machines

This test case was originally written for server testing w/o a time constraint. We should pass an option to stress-ng to keep the execution time to a reasonable value for sru testing.

QA team also reported this problem, pretty sure there's a proposal to add a config option to the p-p-c provider.

Just for clarity, this is already defined in the job description for stress/cpu_stress_ng_test

command:
  if [ -n "$STRESS_NG_CPU_TIME" ]
  then
    echo "Found STRESS_NG_CPU_TIME env var, stress_ng cpu running time is now: $STRESS_NG_CPU_TIME seconds"
    systemd-inhibit stress_ng_test.py cpu --base-time "$STRESS_NG_CPU_TIME"

So as Sylvain suggested, passing an environment variable called STRESS_NG_CPU_TIME set as seconds will alter the default runtime.

hint:
export STRESS_NG_CPU_TIME=600

Also, I did notice that environ: was missing from the job description, so I added it so that a proper environment variable will be honored.

Also, here's the original bug that may not even touch what I was talking about anyway, so feel free to ignore me :D

https://bugs.launchpad.net/plainbox-provider-checkbox/+bug/1815867

Thanks for the fix Jeff!