Metrics & Monitoring

---

Definition

---

Metric Types

// COUNTER: Cumulative value that only goes up
// Use for: Total requests, errors, bytes sent
Counter requestsTotal = Counter.builder("http_requests_total")
    .tag("method", "POST")
    .tag("path", "/orders")
    .register(registry);

requestsTotal.increment();

// GAUGE: Value that can go up or down
// Use for: Current connections, queue size, temperature
Gauge.builder("active_connections", connectionPool, Pool::getActiveCount)
    .register(registry);

// HISTOGRAM: Distribution of values
// Use for: Request latency, response size
Timer requestLatency = Timer.builder("http_request_duration_seconds")
    .publishPercentiles(0.5, 0.95, 0.99)
    .register(registry);

requestLatency.record(() -> {
    handleRequest();
});

// SUMMARY: Similar to histogram, calculated client-side
// Pre-computed percentiles, less server work

// METRIC NAMING CONVENTIONS (Prometheus style)
// - snake_case
// - Include unit: _seconds, _bytes, _total
// - Use _total suffix for counters

// Good names:
http_requests_total
http_request_duration_seconds
db_connections_active
queue_messages_pending

// Bad names:
requestCount
latency
connections

---

Common Metrics

// APPLICATION METRICS

// HTTP Requests
@Around("@annotation(org.springframework.web.bind.annotation.RequestMapping)")
public Object measureEndpoint(ProceedingJoinPoint pjp) throws Throwable {
    Timer.Sample sample = Timer.start(registry);
    String outcome = "success";

    try {
        return pjp.proceed();
    } catch (Exception e) {
        outcome = "error";
        throw e;
    } finally {
        sample.stop(Timer.builder("http_request_duration_seconds")
            .tag("method", getMethod(pjp))
            .tag("path", getPath(pjp))
            .tag("outcome", outcome)
            .register(registry));
    }
}

// Database Queries
Timer.builder("db_query_duration_seconds")
    .tag("query", "findUserById")
    .register(registry)
    .record(() -> userRepository.findById(id));

// External Service Calls
Timer.builder("external_service_duration_seconds")
    .tag("service", "payment-gateway")
    .tag("operation", "charge")
    .register(registry);

// Business Metrics
Counter.builder("orders_created_total")
    .tag("product_type", productType)
    .register(registry).increment();

Gauge.builder("inventory_level", inventory, Inventory::getCount)
    .tag("product_id", productId)
    .register(registry);

// JVM METRICS (usually auto-collected)
// - jvm_memory_used_bytes
// - jvm_gc_pause_seconds
// - jvm_threads_current
// - process_cpu_usage

---

Dashboards & Alerting

---

Monitoring Architecture

---

Tips & Tricks

---

• *