Escalation
Incident Escalation
Escalation is the process of routing an incident to more senior, specialized, or additional responders when the current responder cannot resolve it alone.
Types of Escalation
Hierarchical Escalation: Moving up the chain of command (engineer → senior engineer → manager)
Functional Escalation: Moving to specialists (frontend → database team → security team)
External Escalation: Engaging vendors, partners, or third-party support
Why Escalation Matters
Without clear escalation: - Engineers struggle alone for too long - Incidents drag on unnecessarily - On-call burnout increases - Customer impact extends
Escalation Policy Components
1. Triggers - When should escalation happen? - Time-based: No response in X minutes - Severity-based: SEV1 always escalates - Skill-based: "I don't know this system"
2. Paths - Who gets escalated to? - Primary and backup contacts - Specialty teams - Management chain
3. Methods - How are people reached? - PagerDuty, phone, Slack, SMS
4. Documentation - What info is passed along? - Current status, actions taken, impact
Escalation Best Practices
- No shame in escalating - Better to get help than struggle - Escalate early - Don't wait until you're stuck - Provide context - Brief the next responder effectively - Stay involved - Original responder should assist, not disappear - Review escalation patterns - Too many? Training gap. Too few? Fear of escalating.