An Emergency Playbook is a concise, self-contained guide to formal actions and decisions for urgent situations where bypassing the normal process is unavoidable. In code review, it provides a controlled response to rare events such as a production hotfix, rather than granting permission to skip review whenever convenient. section 10.1 section 10.3
A runbook gives detailed, repeatable instructions for one task, such as rolling back a deployment. A playbook coordinates a broader process, combining multiple tasks with human decisions and approved exceptions. section 10.1
A practical emergency procedure should define:
- The emergency conditions and decision tree that determine whether the normal process may be bypassed. section 10.4
- The small, explicit group authorized to initiate the procedure, adjusted for the risk of each process. section 10.2.2
- The invocation steps, including a ticket or form, justification, documentation, and any required approval. section 10.2.2
- The bypass mechanism, operational actions, communication, follow-up documentation, and postincident analysis. section 10.4
Emergency self-approval during a production incident
A production emergency requires a pull request to move quickly, but the author cannot obtain the usual approver. The playbook allows escalation to next-level managers only during the emergency, preserves the rule against ordinary self-approval, and records the exception. The updated process adds an emergency tag, removes the blocking self-approval setting, and requires a post-approval justification. section 3.4.1
Create the playbook collaboratively with the team and relevant security, DevOps, infrastructure, site reliability, and compliance partners. Make invocation deliberately more tedious than ordinary work, communicate every circumvention, and review the procedure after incidents. section 10.2.2 section 2.4.3
