Avoiding Vanity Metrics
What you'll learn: How to distinguish between metrics that look impressive and metrics that actually drive meaningful decisions in your systems.
The Vanity Trap
Vanity metrics are measurements that make dashboards look busy and numbers trend upward, but don't help you make better decisions. They're like counting steps on a treadmill while ignoring whether you're actually getting healthier—the number goes up, but it doesn't tell you what to do next.
In distributed systems, vanity metrics might include total requests served, total uptime days, or raw CPU usage percentages without context. These numbers feel important, but they don't answer critical questions: Should I scale? Is the user experience degrading? Where should I optimize?
Actionable vs Vanity
Actionable metrics directly inform decisions:
- Error rate (not just error count) → tells you if quality is degrading
- P95 latency (not average response time) → reveals tail latencies affecting users
- Saturation percentage (not raw resource usage) → indicates when to scale
Vanity metrics lack decision context:
- Total requests served → doesn't indicate health or capacity needs
- Uptime in days → ignores recent incidents or degraded performance
- Number of servers → says nothing about efficiency or user experience
The Test
Ask: "If this metric changes, what specific action would I take?" If you can't answer clearly, it's likely vanity. Using the RED method (Rate, Errors, Duration) or Four Golden Signals helps focus on metrics that matter because each directly maps to user impact and system health.
Key Takeaway: Collect metrics that answer "what should we do next?" not metrics that simply show "we're busy." Actionable metrics drive decisions; vanity metrics just occupy dashboard space.