Admin guide¶
This guide is for people who deploy and operate grex: install the binary, image, or Helm chart; configure listeners and TLS; scrape metrics; and understand process lifecycle.
Responsibilities¶
- Choose listen addresses and network exposure for the three listeners
- Deploy on Kubernetes with Helm when running in a cluster
- Configure OpAMP TLS/mTLS for collectors
- Set fleet heartbeat / stale / required attribute policy
- Scrape Prometheus endpoints with safe cardinality limits
- Integrate with orchestrator health probes (
/healthz,/readyz)
Contents¶
| Page | Description |
|---|---|
| Install | Binary, Docker, Compose, Helm overview |
| Releases | Versioned artifacts + changelog |
| Deploy with Helm | Kubernetes install, TLS, Ingress, scrapes |
| Configuration | YAML + GREX_* reference |
| TLS and mTLS | OpAMP server TLS and client certs |
| Authentication | Current state + planned UI/API auth |
| Health and lifecycle | Probes, drain, shutdown |
| Scaling with gateways | OpAMP gateway topology |
| High availability | Target multi-replica architecture (mostly design, not yet shipped) |
Security posture (today)¶
| Surface | Protection today |
|---|---|
| OpAMP listener | Optional TLS + optional mTLS (tls.*) |
| UI / API listener | Open — auth not implemented yet |
| Telemetry listener | Unauthenticated scrape of metrics/health; optional pprof |
Run the UI and telemetry listeners on trusted networks (or behind a gateway you control) until UI authentication ships. Prefer mTLS on OpAMP for any non-lab collector fleet.
Related¶
- Observability — metrics, logs, scrape jobs
- Reference: endpoints
- SPEC — intended 1.0 end state