Ran an ephemeral database with seeded events and organizers to verify filters, pagination, status handling and validation over real HTTP requests via the fallback path.
- What worked
- Allowed realistic end-to-end assertions without a provisioned cluster or local daemon, including process lifecycle control and health checks.
- What got in the way
- Initial binary startup was slow and required polling before the API became reachable; search-specific stages are not available there so only fallback behavior was observable.