Ghost‑CLI In‑Place Upgrade vs Docker Blue‑Green: Which Minimizes Downtime for a Small Site?

Goal

Upgrade a lightweight Ghost CMS instance while keeping the site online.

Constraints

Single‑node deployment on Linux, default SQLite, and a process manager such as systemd or PM2.

Unresolved Decision

Both documented methods claim zero‑downtime, but the real impact on latency and user experience is unclear.

Questions

• Will the brief pause introduced by Ghost‑CLI’s ghost upgrade (which stops and restarts the node process) generate a measurable latency spike that could affect user sessions?
• Does the Docker blue‑green workflow guarantee that traffic is never routed to a shutting‑down container, even when the load balancer requires a brief health‑check period?
• How do the two approaches compare in terms of operational complexity for a small team with limited DevOps resources?