Microservices · Lesson 12

Monitoring and Logging in Microservices

Observing Distributed Services

Monitoring tells you whether a service is healthy and meeting its objectives, while logging explains individual events. In microservices, a request may cross several processes, so correlation and consistent dimensions are essential. Observability should help answer what failed, where, and why.

Recording a Useful Request Event

Structured logs should include stable fields rather than only a sentence. Do not log credentials, tokens, or unnecessary personal information.

import java.time.Instant;

public class RequestLog {
    public static void main(String[] arguments) {
        String traceId = "trace-184";
        String serviceName = "course-service";
        int statusCode = 200;
        System.out.println("time=" + Instant.now()
                + " service=" + serviceName
                + " traceId=" + traceId
                + " status=" + statusCode);
    }
}

Signals and Their Uses

Common Mistake and Practices

Common Mistake: Logging a full request body or stack trace for every normal request. Logs become expensive, noisy, and may expose private data while hiding the events operators need.

• Include a correlation or trace ID.
• Use structured fields and stable names.
• Track latency, error rate, and saturation.
• Define actionable alerts with ownership.
• Redact secrets and sensitive fields.

Real-World Use Case

A learner request receives one trace ID as it passes through the gateway, course service, and database client. A dashboard shows latency and error rates, while linked logs reveal the failing dependency without searching unrelated text.