When all endpoints in a cluster fail health checks, Praxis returns errors to clients. There is no way to route to a fallback backend or serve a static error page during complete outages.
Goal
Configurable backup pools that activate automatically when the primary cluster has no healthy endpoints.
Requirements
- Per-cluster
fallback_cluster configuration referencing another cluster
- Automatic activation when all primary endpoints are unhealthy
- Automatic deactivation when any primary endpoint recovers
- Access log indication when fallback is active
- Works for both HTTP and TCP listeners
When all endpoints in a cluster fail health checks, Praxis returns errors to clients. There is no way to route to a fallback backend or serve a static error page during complete outages.
Goal
Configurable backup pools that activate automatically when the primary cluster has no healthy endpoints.
Requirements
fallback_clusterconfiguration referencing another cluster