Raven
Operations
Deployments, SLOs, and monitoring for Raven
Deployments
- Build from
services/raven/Dockerfile. - The container runs
granianagainstservices.raven.main:app. - Cloud Run is configured for HTTP/2 (h2c) on port 8080.
PORTis injected by Cloud Run; granian should bind to it in production.
Startup behavior
- Initializes Spanner client and warms the pool.
- Loads Bubble API token and OIDC config as needed.
- Initializes Firebase Admin with ADC (used for ID token verification).
Monitoring and alerts
- Logs are structured JSON via
common.logging_utils. - No Sentry integration currently in this service.
SLOs and SLIs
- TBD. Define latency and error targets for search and login RPCs.
Scaling and performance
- Search performance depends on Spanner search indexes.
- Streaming RPCs should be monitored for long-lived connections.
- Search staleness and probe limits can be tuned via settings.
Last updated on