Zero-downtime deployment (blue/green)

v0.7.0-beta

Zero-downtime deployment (blue/green). When deploying an app, the agent now brings up a new replica of the container beside the old one, waits for its health check (HEALTHCHECK) to turn green, and only then removes the old one. During the switchover the reverse proxy spreads traffic across both: your app never goes down. If the new replica's health check fails, the old one keeps serving and the deployment is marked failed (still with no downtime). Backwards compatible: a hub not yet sending this instruction is handled as before (classic recreation).