---
title: Health checks
description: The two endpoints your uptime monitor or load balancer can use to check the instance.
---

Metis Suite has two health endpoints for uptime monitors and load balancers. Both answer with a status code only, with no body, and need no sign-in.

| Endpoint | Answers `200` when | Otherwise |
| --- | --- | --- |
| `/api/livez` | The app process is serving requests. It checks nothing else. | No answer |
| `/api/readyz` | The settings loaded and both databases answer. | `503` |

For example:

```bash
curl -fsS -o /dev/null -w "%{http_code}\n" https://mail.example.com/api/readyz
```

Use `/api/livez` to decide whether to restart the app, and `/api/readyz` to decide whether to send it traffic. The updater also waits for `/api/readyz` after an upgrade.

When `/api/readyz` answers `503`, the reason is written only to the app's log, never to the response:

```bash
sudo docker compose --project-directory /opt/metis-suite logs app
```

## Checks inside the dashboard

The owner dashboard has deeper checks that contact each service: the **Health** card on every driver and provider page under **System**. They run only when you click **Run check**. See [Owner dashboard](/self-hosting/operate/owner-dashboard#health-checks-on-driver-and-provider-pages).