Skip to content

User guide

This guide is for operators who use grex’s web UI to understand collector fleet health. You do not need to run grex yourself—someone may have deployed it for you—but you need a browser and the UI URL.

What you can do

  • Browse all known collectors (agents and gateways) on the fleet page
  • Filter by health, connection, gateway path, and attributes
  • Open an agent detail page for attributes, capabilities, and effective config
  • View server status (grex version, uptime, fleet summary counts)

What you cannot do (1.0)

grex is read-only. The UI does not:

  • Edit or push remote configuration
  • Restart collectors
  • Upgrade agent packages
  • Change grex server settings

Those capabilities are out of scope for 1.0 (see SPEC).

Access and authentication

Authentication not yet available

UI and API login (mTLS client certs and OIDC via Dex) is planned but not implemented. Today, anyone who can reach the UI listener can use the fleet views. Operators should treat the URL as internal until auth ships.

Pages

Page Path Description
Fleet overview / Table of agents with filters and auto-refresh
Agent detail /agents/{instance_uid} Full snapshot for one collector
Server status /status grex build info and fleet counts

Concepts in one minute

Term Meaning in the UI
Connected Agent checked in recently enough (or TCP still open for direct connections)
Disconnected Missed a check-in window but not yet evicted; last health is retained
Healthy / Unhealthy From the agent’s OpAMP health report (not the same as connected)
Via gateway Session is multiplexed through an OpAMP gateway collector
Role Best-effort label: service.component, or Gateway if service.name contains “gateway”, else Collector

More detail: Fleet state (developer).