A team is spending excessive runner time on jobs that wait indefinitely for a failed external service. Which workflow control directly bounds each affected job execution time?

1. Set a concurrency group but do not enable any timeout or cancellation behavior.
2. Set matrix max-parallel: 1 and assume each running job receives a 15-minute execution deadline.
3. Set retention-days on the uploaded logs and expect the job to stop at that duration.
4. Configure job timeout-minutes alongside application-level timeouts.