Skip to main content

Usage statistics & metrics

The Go SDK can optionally batch and send:

  • usage statistics (feature checks + usage)
  • metrics (measurements / counters / observations)

This is done via gRPC.

Enable usage statistics​

client, err := toggly.NewClient(toggly.Config{
  AppKey:       "YOUR_APP_KEY",
  Environment:  "Production",
  EnableUsage:  true,
})

Record usage​

Checks are automatically recorded when EnableUsage=true.

RecordUsage is count-only. For attempt plus duration and errors, use a scope:

ctx := toggly.Context{Identity: "user-123"}
err := client.Use(context.Background(), "MyFeature", ctx, func(scope usage.FeatureUsageScope) error {
    return processOrder()
})
// Pass context.Canceled (or a wrapped cancel) from fn to record attempt only.

Bare “used” events without timing:

client.RecordUsage("MyFeature", true, toggly.Context{Identity: "user-123"})

See Feature Usage Tracking.

Enable metrics​

client, err := toggly.NewClient(toggly.Config{
  AppKey:        "YOUR_APP_KEY",
  Environment:   "Production",
  EnableMetrics: true,
})

Emit metrics​

The final argument is the variant name; use an empty string for a boolean feature.

m := client.MetricsClient()
if m != nil {
  featureKey := "MyFeature"
  m.Increment("MyFeature.Counter", 1, &featureKey, "")
  m.Measure("MyFeature.LatencyMs", 12.3, &featureKey, "")
  m.Observe("MyFeature.PayloadSize", 2048, &featureKey, "")
}

Flushing​

Both usage and metrics are flushed periodically (defaults to 1 minute). You can set:

  • UsageFlushInterval
  • MetricsFlushInterval

The underlying clients also expose a Flush(ctx) method if you need a manual flush.