Operations
The Operations page (/dashboard/operations) gives SRE-style signals for a project: SLO-style availability, latency, spend forecast, provider health overlay, and a lightweight incident timeline.
Metrics
| Card | Description |
|---|---|
| SLO availability | Share of successful requests vs target (default target 99.9%) |
| P95 latency | 95th percentile latency for the selected range |
| LLM spend forecast | Projects 30-day LLM spend from current run rate |
| Provider health | Count of providers marked healthy vs degraded |
Environment filter
Use the global environment filter in the dashboard header (production, staging, development, or all). Operations metrics respect the same filter as Overview and Providers.
Provider status overlay
Per-provider cards show Healthy or Degraded based on recent error rate thresholds.
Incident timeline
Automated entries (retry storms, provider outages, cost spikes, etc.) appear when background analysis detects issues. Pair with Alerts for notifications.
API
GET /api/ops/metrics?projectId=&range=&environment= (Clerk JWT required).