salvo-graceful-shutdown

Graceful shutdown in a Salvo web server is implemented by registering a listener that triggers ServerHandle::stop_graceful upon receiving termination signals. This mechanism waits for in-flight requests to complete before stopping the server, ensuring zero-downtime updates.

Graceful shutdown is the process of finishing in-flight requests before a server stops, preventing abrupt connection drops and data inconsistency. You need it for production web servers to maintain data integrity during deployments, Kubernetes pod terminations, or manual restarts.

Yes, Salvo supports cross-platform signal handling for Kubernetes lifecycle events by integrating with tokio::signal. It captures Ctrl+C and SIGTERM signals to initiate a timed graceful stop, coordinating readiness checks with load balancers and orchestrators.

You run cleanup tasks before a Salvo server stops by registering them within the shutdown listener logic. When a termination signal is received, the server executes these optional cleanup tasks to release resources before finalizing the graceful shutdown process.

The best way to handle SIGTERM and readiness checks with Salvo and Tokio is using ServerHandle::stop_graceful. It intercepts cross-platform signals via tokio::signal, cooperates with load balancers through health-check integration, and waits for a timed graceful stop.