Platform Health
Monitor infrastructure metrics for API, database, Redis, and host resources.
Platform (/platform) shows infrastructure health for administrators. Requires platform.read (ADMIN by default). Users without this permission are redirected to the Operations Center.
What you'll see
| Area | Metrics |
|---|---|
| API | Process memory, heap, memory limit and utilization |
| Redis | Memory usage, key count, live latest-value cache size |
| PostgreSQL | Database size, telemetry table size, connection count vs max |
| Host / disk | Disk usage, host memory utilization |
| Tables | Top tables by size and approximate row counts |
| Telemetry queue | Pending JetStream messages not yet persisted — rising values may indicate worker capacity pressure |
Metrics refresh on load; use for capacity planning and incident triage.
When to use
- After deployment — verify API and database are within expected bounds
- During incidents — identify memory pressure or connection exhaustion
- Capacity planning — monitor telemetry table growth over time
Operations Center vs Platform
| Page | Focus | Permission |
|---|---|---|
Operations Center (/operations) | Operational KPIs: devices, rules, alarms, EPS, alarm/telemetry feeds | ops.read |
Platform (/platform) | Full infrastructure detail and table statistics | platform.read |
Operations Center includes a summary row of platform metrics when you have platform.read; Platform is the detailed drill-down.
Health checks for load balancers use GET /api/health (no authentication). Deeper metrics remain behind platform.read.
Related API
- System — health endpoint
- Operations — operational KPI endpoints