AI Gateway feature
Observability
Available
Understand production AI traffic end-to-end: latency, errors, usage, and cost by model/provider/route, so you can debug, govern, and optimize with confidence.
Capabilities
- Latency, error, usage, and cost metrics
- Activity logs with configurable retention (by plan)
- Export hooks (as capabilities mature)
- Route-level insight to compare providers and models
Common use cases
- Identify which provider is causing incidents and failover faster
- Track budget drift for a feature and cap spend per environment
- Compare model quality/cost tradeoffs route-by-route
- Audit usage for internal chargeback and procurement
Faster debugging
See where errors come from and which provider/model/regional route is failing.
Cost attribution you can trust
Understand spend by team, environment, model, and workload.
Optimization loop
Use signals to refine routing policies and reduce waste over time.
FAQ
Answers reflect current direction and may evolve as the platform ships.
