Custom Metrics

Overview

Custom metrics represent business or domain-specific behavior that default infrastructure metrics cannot capture.

When to Add Custom Metrics

• Critical workflow steps need visibility
• Business impact must be measurable
• Default host metrics miss service-level behavior

flowchart LR
    A[Business Event] --> B[Application Metric]
    B --> C[Monitoring Backend]
    C --> D[Alert and Dashboard]

Practical Example

OrdersCreatedTotal
    .WithLabels("eu-west-2", "web")
    .Inc();

PaymentFailuresTotal
    .WithLabels("timeout")
    .Inc();

Production Example

Subscription platforms monitor trial conversion and renewal failures as first-class observability signals.

Common Pitfalls

• Creating metrics with excessive label cardinality
• Defining metrics without clear ownership
• Shipping metrics that are never used by dashboards or alerts

Key Takeaways

• Custom metrics bridge technical telemetry and business outcomes.
• Clear naming and bounded labels prevent scale issues.
• Metrics should map directly to decisions and actions.