Fix job timeout leaving VM behind

What does this MR do?

Adds an internal timeout so we don't reach the top-level job timeout and leave VMs behind by skipping the after_script.

In order to make use of the timeout command, move the script to a bash script. Added some further refactoring in the following commit once it was in bash.

Why was this MR needed?

Jobs are leaving VMs behind

What's the best way to test this MR?

Forced the job to timeout in https://gitlab.com/gitlab-org/ci-cd/shared-runners/images/macstadium/orka/-/jobs/1052847214. You can see that the after_script has been executed.

What are the relevant issue numbers?

Closes #62 (closed)