Disposability [principle] v1.0.0

Processes are disposable — they can be started or stopped at a moment's notice. Fast startup makes scaling and deploys responsive; graceful shutdown on SIGTERM (stop accepting new work, finish in-flight work, exit) makes restarts safe. Crash-only design is the goal: a kill -9 should not corrupt state.

Start fast. Shut down gracefully. Tolerate sudden death.

Applies To

• Cold-start time measured in seconds, not minutes — measure and budget it.
• On SIGTERM: stop accepting new requests, drain the queue, close DB connections, exit.
• Workers must use idempotent jobs or atomic transactions so a mid-job kill is safe.

Disposability [principle] v1.0.0

Processes are disposable — they can be started or stopped at a moment's notice. Fast startup makes scaling and deploys responsive; graceful shutdown on SIGTERM (stop accepting new work, finish in-flight work, exit) makes restarts safe. Crash-only design is the goal: a kill -9 should not corrupt state.

Start fast. Shut down gracefully. Tolerate sudden death.

Applies To

• Cold-start time measured in seconds, not minutes — measure and budget it.
• On SIGTERM: stop accepting new requests, drain the queue, close DB connections, exit.
• Workers must use idempotent jobs or atomic transactions so a mid-job kill is safe.

Relations

enhances: [@backend-architecture/principle-concurrency-process-model]

Enhances

• @backend-architecture/principle-concurrency-process-model