Using Structured Logging

Structured logging is a method of logging where log entries are formatted in a way that machines can easily parse them. Using structured logging enables better log analysis and easier querying of logs. Go provides built-in support for structured logging through the log/slog package, introduced in Go 1.21.

Basic Structured Logging with slog

Below is an example demonstrating how to use the slog package for structured logging.

package main

import (
	"log/slog"
	"os"
)

func main() {
	// Initialize default logger
	logger := slog.New(slog.NewTextHandler(os.Stdout, nil))
	slog.SetDefault(logger)

	// Structured logging
	slog.Info("User logged in",
		"event", "user_login",
		"user", "johndoe",
	)

	// Another structured log entry
	slog.Warn("File upload completed with warnings",
		"event", "file_upload",
		"user", "johndoe",
		"size", 3456,
	)
}

Output Example

When you run the previous code, the structured logs will be in JSON format by default:

time=2009-11-10T23:00:00.000Z level=INFO msg="User logged in" event=user_login user=johndoe
time=2009-11-10T23:00:00.000Z level=WARN msg="File upload completed with warnings" event=file_upload user=johndoe size=3456

Best Practices

• Use a structured logging library, like logrus, to avoid manually formatting JSON.
• Consistently use fields to capture important context about log events (e.g., user IDs, operation types).
• Set log levels appropriately to avoid unnecessary verbosity in production environments.

Common Pitfalls

• Forgetting to set the output of logs, resulting in logs being directed to stderr where they might be missed or ignored.
• Using plain log messages without any structured fields can make searching and filtering difficult.
• Ignoring performance considerations when logging extensively in high-throughput applications.

Performance Tips

• Buffer log output to minimize impact on application performance.
• Avoid excessive logging in hot paths to prevent bottlenecks.
• Consider asynchronous logging to reduce latency.