Describe the feature
Support running multiple Kival server instances against the same PostgreSQL database and shared blob storage behind a load balancer.
Kival currently targets a single application instance. The implementation should audit process-local assumptions and ensure correctness does not depend on requests reaching the same server process.
Areas to review include authentication/session flows, realtime notifications and invalidation, background work, in-memory coordination, startup behavior, and attachment storage.
PostgreSQL should remain the authoritative shared state where possible, and multi-instance operation should not require sticky sessions.
Add integration coverage that exercises important behavior across at least two independent Kival server instances.
Additional context
Shared blob storage is a prerequisite and is tracked separately.
Kubernetes-specific deployment tooling, autoscaling, and distributed caching are out of scope unless required for correctness.
Describe the feature
Support running multiple Kival server instances against the same PostgreSQL database and shared blob storage behind a load balancer.
Kival currently targets a single application instance. The implementation should audit process-local assumptions and ensure correctness does not depend on requests reaching the same server process.
Areas to review include authentication/session flows, realtime notifications and invalidation, background work, in-memory coordination, startup behavior, and attachment storage.
PostgreSQL should remain the authoritative shared state where possible, and multi-instance operation should not require sticky sessions.
Add integration coverage that exercises important behavior across at least two independent Kival server instances.
Additional context
Shared blob storage is a prerequisite and is tracked separately.
Kubernetes-specific deployment tooling, autoscaling, and distributed caching are out of scope unless required for correctness.