golang-samber-slog

To build a structured logging pipeline in Go using slog, combine sampling, formatting, routing, and sink handlers via slog-multi. This ensures logs are routed by severity, sampled for noise control, and safely flushed during graceful shutdown.

slog handler composition chains multiple handlers so you can sample noisy records first, scrub sensitive PII fields, and then route logs by severity to different backends like Sentry or Loki, keeping downstream handlers clean and efficient.

Yes, you can integrate slog HTTP middleware with Gin, Echo, Fiber, Chi, or standard net/http. These integrations provide request logging and automatically attach request-scoped attributes to your structured logs.

The best way to route Go application logs to Datadog, Loki, and Sentry is using samber/slog with slog-multi to route records by severity. Errors go to alerting backends, while info logs go to observability storage.

Buffered slog handlers lose logs during server shutdown because they require a graceful shutdown process to flush pending records. You must explicitly close the handlers to ensure all buffered logs are delivered to their backends before exit.

Yes, slog supports bridging logs from existing Zap, Zerolog, and Logrus implementations. This allows you to migrate to a unified structured logging pipeline without rewriting all existing logging calls immediately.