Metrics & telemetry

See what your workloads are doing — scrape the metrics they already expose, or switch on telemetry and send traces, metrics and logs with zero code changes.

Scraping metrics

If a component exposes a metrics endpoint (the Prometheus text format most frameworks emit out of the box), tell High the port and path in its config. The platform collects it from the outpost — including outposts behind NAT and managed clusters, which push their pages home over the private overlay. No public scrape port, no firewall holes.

Workload telemetry

Switch telemetry on for a component and High injects the standard OpenTelemetry environment into it — endpoint, headers and a per-stage token are all set for you. Any app or SDK that speaks OpenTelemetry starts delivering traces, metrics and logs on its next deploy, without code changes.

The token ties everything to its stage, so one workload can never write into another's data — and revoking a stage's token cuts it off immediately.

Where to look

The project's metrics view charts what your components send. Platform admins additionally get a usage dashboard per organization. Compute usage itself is metered separately — see rent & earn for how consumption turns into statements.