MTTI
Mean Time To Identify
MTTI (Mean Time To Identify) measures the average time from incident detection to understanding what's causing the problem.
The Identification Phase
After you know there's a problem (detection), you need to understand: - What's actually broken? - What's the root cause? - What's the impact scope?
This is the identification phase, measured by MTTI.
Why MTTI Often Dominates
For many teams, MTTI is where most time is lost:
| Phase | Typical Time | % of Total |
|---|---|---|
| Detection | 5-15 min | 10-20% |
| **Identification** | **15-45 min** | **40-60%** |
| Resolution | 10-30 min | 20-30% |
The actual fix is often quick once you know what's wrong!
Why Identification Takes So Long
- Context is scattered across 10+ tools - No correlation between related events - Tribal knowledge lives in people's heads - Poor observability into system behavior
How to Reduce MTTI
1. Centralize operational context - Everything in one searchable place 2. Automate correlation - Link deployments → errors → incidents 3. Document system knowledge - Runbooks, architecture diagrams 4. Improve observability - Logs, metrics, traces that tell the story 5. AI-assisted analysis - Surface likely causes automatically