Describe in detail what your merge request does and why.
When tracing a pipeline using a job that runs in the same pipeline, we run into the problem that the pipeline has not completed yet since the trace job is still running. This necessitates some workarounds to get the finished_at timestamp of the pipeline.
This MR changes the tracing job to run in a child pipeline. The parent pipeline (coordinator pipeline) will trigger the child pipeline but won't wait for it to complete. This results in the coordinator pipeline having completed, by the time the tracing job is run.