What metrics will you use to evaluate the performance and health of a service?

To assess the performance and health of a Go service, I usually use the following metrics:

• Response Time (Latency) — how long it takes to process a request.
• Throughput — the number of requests processed per unit of time.
• CPU and memory usage — resource monitoring to avoid overloads.
• Number of goroutines — too many goroutines may indicate leaks or management issues.
• Errors and exceptions — the number of errors to monitor stability.
• GC (Garbage Collector) metrics — frequency and duration of garbage collection, affecting performance.

To collect these metrics, I often use the built-in pprof package and third-party monitoring systems, such as Prometheus with Grafana for visualization.