feat(bigtable): add Cloud Bigtable client uptime gauge metric

[mgarolera]

Add reporting for CBT client uptime to help with observability and troubleshooting

[gemini-code-assist]

Code Review

This pull request introduces a new client uptime metric (uptime) as an Int64ObservableGauge and adds support for exporting Gauge[int64] and Gauge[float64] metrics to Cloud Monitoring. Feedback suggests ensuring that the StartTime of the TimeInterval is set equal to the EndTime for gauge metrics to comply with Cloud Monitoring API requirements. Additionally, a defensive check should be added to ensure tf.startTime is initialized before observing the uptime metric.

[mutianf]

I think this needs a bit more work. All the metrics exported from this exporter goes to bigtable_client_raw which maps to BigtableTable target instead of BigtableClient target, which is different from the internal metric definition.

[sushanb]

This metric should be exported similar to client startup time.

google-cloud-go/bigtable/internal/transport/connpool.go

Line 636 in f8eb671

clientStartupTime, err := meter.Float64Histogram(

this will use the cloud.BigtableClient schema.

[mutianf]

Ah yes. I think we should move the definition of this metric to otel_metrics, something like:

  otelMeterProvider := metric.NewMeterProvider(meterOpts...) // existing code
  startTime := time.Now()
   _, err := meter.Int64ObservableGauge(
      "uptime",
      metric.WithDescription("Time since the Bigtable client was created."),
      metric.WithUnit("s"),
      metric.WithInt64Callback(func(_ context.Context, o metric.Int64Observer) error {
          o.Observe(int64(time.Since(startTime).Seconds()))
          return nil
      }),
  )