Upstream Health
Upstream health monitoring tracks whether each route's backend is responding correctly.
What it does
ProxyOS runs active health checks on a per-route configurable schedule. Results are stored in the health_checks table and contribute to a composite health score (route_health_scores) that combines uptime percentage, p95 latency, and SLO compliance.
The health score is shown on the routes list as a colored indicator and as a detailed chart on the route detail page.
When to use it
Health monitoring is enabled by default on all routes. The health score feeds into:
- Dashboard route list (quick visual status)
- Alert rules (trigger notifications when a route goes unhealthy)
- Load balancing (remove unhealthy upstreams from rotation)
- SLO compliance tracking
How to configure
Health check settings are on the route settings page. See Health Checks for detailed configuration options.
Alert rules: Create alert rules in Settings → Alerts to send notifications when a route's health score drops below a threshold or when a health check fails for N consecutive intervals. Alert events are stored in alert_events.
Troubleshooting
- Health check red but service accessible: see Upstream Health Failed
- Health score not updating: check that health check is enabled on the route and the interval has elapsed since the last check