go-observability

Instrument Go services by adding Prometheus counters and histograms for request rates and latency, and create OpenTelemetry spans for database operations and outbound API calls with proper context propagation to make production issues visible.

Signal correlation connects metrics to traces using Prometheus exemplars and propagates trace IDs, allowing on-call debugging to move from a Grafana alert directly to the slow trace and identify the exact database call causing the spike.

Yes, this instruments HTTP handlers with bounded Prometheus metrics for latency and error monitoring, while recording OpenTelemetry error statuses to ensure production debugging replaces guesswork with observable request paths.

Add trace correlation to database operations by creating OpenTelemetry spans around DB calls with proper context propagation, linking them to Prometheus histograms via exemplars so slow queries are traceable from metric alerts.

Yes, bounded labels are required for Prometheus counters and histograms to monitor request rates, error rates, and latency without unbounded cardinality, ensuring metrics remain performant and traceable alongside OpenTelemetry spans.

Exemplar support is needed to link Prometheus metric data points directly to OpenTelemetry traces, enabling fast root cause identification by jumping from latency or error alerts in Grafana to the specific request trace.