# The Monitoring page

Where you connect monitoring tools from the Platform console, and see what each process sent.

In **Platform**, under **Settings**, open **Monitoring**. Every platform admin can read it. Only an instance-wide platform admin changes it, and a save asks for your [two-factor code](https://docs.gorunner.io/self-host/platform#two-factor-codes).

![Where monitoring goes: Prometheus Off, OpenTelemetry and Sentry Not connected, Logs Live, each with Set up or Edit.](https://docs.gorunner.io/img/self-host/monitoring-light.jpg)

## Where monitoring goes

One row per tool: **Prometheus**, **OpenTelemetry**, **Sentry** and **Logs**, with its **State** and **Last sent**. **Set up** or **Edit** opens its panel. Nothing is sent anywhere until a tool is connected.

- **OpenTelemetry**: **Endpoint**, **Protocol**, **Headers**, **Service name**, **Traces kept** (a percentage), and a switch each for traces, metrics and logs.
- **Prometheus**: **Serve /metrics**, its **Token**, **Split by workflow**, and a **Scrape config** naming the processes reporting now.
- **Sentry**: **DSN** and **Environment**. **Send test error** sends a harmless error from the API.
- **Logs**: The **Level**. The format is set in `.env`.

A saved change reaches every process within seconds, with no restart. **Send test** sends one item through the connection as saved, and says whether it arrived: “Test trace arrived · 2 ms · 18:26”. It asks for no code.

## Each process

What each API and worker process sent in its last report, one column per connection: **Sent** some seconds ago, **Ready**, **Failing** or **Off**. Processes report every 30 seconds.

Under a failing connection, one line says which process, since when, what went wrong and how much was dropped: “Your tool refused the export: 401 Unauthorized.” Runs are never affected.

## Start-up settings

Settings read from the server’s environment when a process starts: database and Redis spans, trace ids on calls steps make, the metric export interval, extra resource attributes, whether OpenTelemetry is allowed, browser error reports, how long the audit log and credential use are kept, and event streaming. Change them in `.env` and restart.

## Settings in .env

A setting also in `.env` shows locked: “Set in .env as …, which wins over this screen. Change it there and restart.” Secrets are never shown back, only their last four characters; leave a field blank to keep it.

**How does the metrics token work?**

**Make a token…** (or **Rotate…**) asks for your code, then shows the token once, starting `grm_`, with your scrape config. “You won’t see this token again.”

Rotating stops the old token at once: update Prometheus before its next scrape. **Remove…** makes `/metrics` answer `503`.

**What if the installation is paused?**

The Platform console, this page included, is closed until a license is loaded. Monitoring set in `.env` keeps sending.

## Related

- [Connect a monitoring tool](https://docs.gorunner.io/self-host/monitoring-tools): Each tool’s address and key · 10 min
- [The Health page](https://docs.gorunner.io/self-host/health): Services, queues and workers, live · 4 min
- [Metrics](https://docs.gorunner.io/self-host/metrics): Every metric

Source: https://docs.gorunner.io/self-host/monitoring-page