rolling restart
Replacing the processes or instances of a service one at a time, starting each new one before stopping an old one, so the service never goes down.
In more detail
Waiting for the replacement to report ready before retiring the old one is what keeps capacity up. Each old process should drain its in-flight requests before it exits.
Related terms
- zero-downtime deployment — Releasing a new version while users keep being served, with no period where the service is unavailable.
- connection draining — Stopping new traffic to a backend while letting its existing connections finish, before removing or restarting it.
- rolling update — Replacing instances of an application with the new version a few at a time while the rest keep serving.
Taught in
- Cluster Architecture: node:cluster & Zero-Downtime — Node.js Internals & Backend Services