Analytics
Under Analytics, review gateway operations for cost reviews, capacity planning, and incident triage. Metrics come from real Proxy traffic (Chat, Images, Tools). Playground probes do not feed this usage pipeline.
Typical views (labels follow the UI):
| View | Use |
|---|---|
| Models | Requests, tokens, and spend by model |
| Providers | Usage and cost by upstream vendor |
| Users | Consumption and activity by user / customer |
| Reliability | Failure rates, latency, availability signals |
What to do in Admin
Section titled “What to do in Admin”- Open the Analytics sub-page you need.
- Pick a time range; narrow by model, provider, or user when useful.
- Cross-check spikes or failure clusters in Request logs.
- Tune route priority, backup providers, budgets, or the Tools Active engine based on what you find.
How it relates to logs
Section titled “How it relates to logs”| Page | Focus |
|---|---|
| Analytics | Aggregated trends and comparisons |
| Request Logs | Per-call detail and pricing summaries |
| Audit Logs | Budget charges and user / key lifecycle |