prometheus-metrics-specialist

Instrument services by exposing a /metrics endpoint and applying standard naming conventions to Counter, Histogram, and Gauge metrics. Add metrics middleware to capture HTTP endpoints, database queries, Kafka events, and business logic, then write PromQL queries to validate dashboards and alerts.

Use service-prefixed operational metrics for infrastructure (latency, error rates, throughput) and separate business-metrics patterns for domain logic. Follow consistent label design and naming schemes to enable cross-service aggregation and readable PromQL queries.

PromQL queries calculate error rates, latency percentiles, and troubleshooting patterns from your exposed metrics. Start with basic aggregation queries on Counter, Histogram, and Gauge data, then refine for dashboard visualization and alert thresholds.

Yes. Instrument Kafka event processing, LLM API calls, batch jobs, and database queries alongside HTTP endpoints using the same middleware and naming patterns, ensuring consistent service- and business-metrics instrumentation across all workload types.

Metrics middleware intercepts HTTP endpoints, database queries, and other operations to automatically collect Counter, Histogram, and Gauge data. Configure it to expose aggregated metrics at /metrics and align with your service's naming and label design.