Course contentsShow
← Cloud and Infrastructure
Lesson 2102 of 2,98945. Observability - Dashboards and SLOsFree lesson

Metric Selection and Relevance

Choose meaningful metrics that align with business goals and avoid vanity metrics that don't drive action.

Metric Selection and Relevance

What You'll Learn

You'll learn how to choose metrics that drive real decisions instead of cluttering your dashboards with numbers that look impressive but don't help you act.

The Problem with Vanity Metrics

When building dashboards (which you've just learned to structure properly), it's tempting to display every available metric. But more data doesn't mean more insight. A dashboard filled with numbers that rise and fall without connecting to business goals becomes wallpaper—visually interesting but ultimately ignored.

Meaningful vs. Vanity Metrics

Meaningful metrics answer critical questions:

  • "Can users complete transactions?"
  • "Are we meeting our SLA commitments?"
  • "Where should we invest optimization effort?"

Vanity metrics look good but don't drive action:

  • Total page views (without conversion context)
  • Raw CPU percentage (without understanding impact)
  • Request counts (without success rates or latency context)

The Actionability Test

Ask yourself: "If this metric changes, what would I do differently?"

If you can't answer clearly, the metric probably doesn't belong on your operational dashboard. It might be useful for a specialized report, but shouldn't occupy precious dashboard real estate.

Aligning with Business Goals

Your infrastructure exists to serve applications that serve business outcomes. Trace backwards:

  • Business goal: "Users can check out successfully"
  • Application requirement: "Payment API responds within 2 seconds"
  • Infrastructure metric: "Payment service p95 latency" + "Payment service error rate"

This connection ensures your dashboard tells a story that matters to stakeholders beyond your team.

Context Over Raw Numbers

A metric alone is just a number. Add context through:

  • Thresholds showing acceptable ranges
  • Trends revealing patterns over time
  • Ratios like error rates instead of raw error counts

Key Takeaway: Choose metrics you can act on that align with business outcomes—if a number going up or down doesn't change your next decision, leave it off the dashboard.