Operations
Monitoring
Track requests, token spend, latency, and errors per project from the dashboard or via the usage API.
Dashboard
The Usage panel shows aggregated metrics for each project:
- Requests — total gateway calls in the selected window
- Tokens — input + output token totals
- Average latency — p50 gateway round-trip in milliseconds
- Error rate — percentage of requests with non-successful finish reasons
The Audit log shows individual request outcomes, filterable by provider, outcome, and project.
Querying usage programmatically
curl — tokens in last 24h
SINCE=$(node -e "console.log(Date.now() - 86400000)")
curl "https://data.plurence.com/v1/usage?timestamp=ge=$SINCE&limit=1000&order_desc=timestamp" \
-H "Authorization: Bearer YOUR_JWT" | \
jq '[.[] | .tokens.total // 0] | add'Alerting Coming soon
Configure threshold-based alerts on token spend, error rate, and latency. Notifications via email and webhook.
Metrics export Coming soon
Export usage metrics to Prometheus, Datadog, or any OpenTelemetry-compatible collector.