An active-passive configuration runs one environment as the primary service for user requests while keeping a duplicate environment idle as backup. Failover redirects traffic to the passive environment when the active one fails or a change causes trouble. section 9.1.2
The pattern reduces the blast radius of infrastructure changes and improves reliability, but it requires duplicated resources and a tested traffic-switching path. Infrastructure as code can reproduce the active region in a passive region. section 9.1.2 section 9.4
A practical implementation follows these steps:
- Duplicate the production environment in another region, changing its region, network range, names, labels, and version identifiers. section 9.1.2
- Keep normal traffic on the active environment while testing changes there. section 9.1.2
- If the active environment fails, assign traffic to the passive environment; otherwise, update the passive environment after validating the change. section 9.1.2
Regional sundew fallback
A production sundew system serves from us-central1 while an idle duplicate runs in us-west1. A load-balancing configuration gives the active service weight 0 during failover and the passive service weight 100, temporarily sending all requests to the backup. section 9.1.2
