telemetry_metrics

Metrics

Prometheus

Counters

• Simple, increment only metrics that keep track of the number of occurrences of a specific event or activity
• Examples:
  • API requests
  • error occurrences
  • system restarts
• When to use:
  • record value that only goes up
  • want to calculate the rate of increase
• Use case:
  • Capacity Planning

Gauges

• Provide snapshot of the particular value at a specific point in time
• Examples:
  • CPU Usage
  • Memory Usage
  • Number of active connections
• When to use:
  • Record value that goes up and down
  • Don’t need to calculate rate
• Use case:
  • Performance Optimization

Histograms

• Measure distribution and frequency of time durations for specific events
• Examples:
  • Request Duration
• When to use:
  • Want to take many measurements of the values and later calculate average or percentiles
  • Not bothered about the exact values and are happy with an approximation
  • When you know the range of values upfront, so you can use the default bucket definitions or define your own
• Use case:
  • User experience enhancement

Timers

Summary

Google’s 4 Golden Signals

• Errors
• Latency
• Throughput
• Saturation