Does anyone know what the first digits of the jobgroup that are shown in the spark ui mean when using all purpose clusters to launch multiple jobs?

Right now the pattern is something like: [id_random]_job_[jod_id]_run-[run_id]_action_[action].