Why we rehearse journald breadcrumbs before touching Ansible handlers

Bare-metal releases fail quietly when handlers restart services without leaving enough context in journald. Our playbook rubric forces students to annotate each handler with a deterministic unit name and a marker line that appears even when the task is skipped. The second paragraph covers ordering traps: notifying multiple handlers without explicit listen segments creates race conditions that Ansible will not warn you about. We map those races on a whiteboard before typing YAML. Third, we explore partial failures: when a health check passes but the business metric lags, journald alone will mislead you. We pair log markers with a minimal HTTP probe recorded in the same journal slice. The closing paragraph documents rollback rehearsals: students narrate a recording while rolling backward, proving they can correlate timestamps with operator voice notes. It is theatrical but effective.