Stage 3 · Build
Production-Ready Services
Structured Logging
Emit JSON logs with slog or zap, attach request IDs, fields, levels, and error context.
7 min readBuilding Backend ApplicationsBuild
Why Structured Logs
Unstructured log lines like 'user logged in' are hard to search and parse. Structured logs emit JSON with named fields. This enables filtering, alerting, and analytics on log data.
log/slog
Gosetting-up-slog.go
18 linesLn 1, Col 1Go
Go 1.21+ includes log/slog in the standard library. NewJSONHandler outputs JSON. Set it as the default logger. All subsequent slog.Info/Debug/Error calls output structured JSON.
Fields and Context
Goadding-fields-to-logs.go
22 linesLn 1, Col 1Go
slog.With creates a logger with pre-attached fields. slog.WithContext attaches it to the context. Every log entry in the handler includes these fields automatically.
Log Levels
| Level | Use Case | Example |
|---|---|---|
| Debug | Detailed diagnostic information | SQL query, request headers |
| Info | Normal operation events | Request completed, job finished |
| Warn | Potential problems | Slow query, high memory usage |
| Error | Failures requiring attention | Database connection failed, job failed |
Error Logging
Gologging-errors-with-context.go
27 linesLn 1, Col 1Go
Always include the error, the entity ID, and any relevant context. Wrap errors with fmt.Errorf for the call stack. Log at the point of failure, not higher up.
Log Aggregation
- stdout/stderr: Docker and Kubernetes collect container logs.
- File-based: Use log rotation with lumberjack or similar.
- Centralized: Ship to Datadog, ELK, or CloudWatch.
- Structured JSON: Enables parsing and querying.
Log to stdout in containers
In containerized environments, always log to stdout. Docker and Kubernetes capture stdout automatically. Do not write to files inside containers.
Mark this lesson complete to store local progress and unlock a cleaner resume path the next time you visit.