# Health Checks

> What Health Checks look for in your setup, what each finding means, what to do about it, and how to stop flagging a finding you accept.

Source: https://proxus.io/docs/organize/health-checks

**Health Checks** look at every configured object (gateways, devices, tags, rules, functions, pipelines, outbound
targets and dashboards) and how they connect, and tell you what in the setup is broken, drifting or wasted. It is
in **Organize > Health Checks**.

Where the [Operations Center](https://proxus.io/docs/monitor/operations-center) tells you what is wrong right now, Health Checks also
tell you what follows from it: a device that stopped means a rule that cannot fire and a target that gets nothing.

The checks run when you open the page. **Check again** runs them again. The line under the title says how many
objects were checked and when. **Open Topology** opens the [Topology](https://proxus.io/docs/monitor/topology) map.

## Findings

The **Findings** tab starts with a one-line verdict, then four cards and the list.

*Screenshot: Health Checks with the four cards Broken, Needs a look, Waste and Passed, a finding about one gateway carrying everything, and the selected finding's connections on the right*

1. The cards. Select one to show only its findings; select **Passed** to show all of them again.
2. A finding: what is wrong, the object, why it matters, what it affects and **What to do**.
3. **Don't flag again** stops flagging this finding for this object. It is offered for everything except broken
   findings.
4. The selected finding: the chain of objects from the source to what is affected. For a broken finding it is titled
   **Where the data stops**, and the broken link is marked.

| Card | Meaning |
| --- | --- |
| **Broken** | Data or automation has stopped. |
| **Needs a look** | The edge runs something different from what is configured here, or the setup carries a risk. |
| **Waste** | Data is collected or computed and then used by nothing. |
| **Passed** | Checks with nothing to report. The list of passed checks is under **… checks passed** at the bottom. |

Each finding has a button that opens the object (**Open device**, **Open rule**, **Open gateway** and so on).
**Trace this object**, in the panel on the right, opens it in the **Trace an object** tab.

## The checks

| Finding | Card | What it means | What to do |
| --- | --- | --- | --- |
| **Gateway is offline** | Broken | The hub has no connection to the gateway. Devices, rules and functions running on it are stopped. | Check that the gateway host is running and can reach the hub. |
| **Device stopped sending data** | Broken | An enabled device sent data before, but nothing arrived within the expected time. | Check the device's network and connection settings, or disable the device if it was retired. |
| **Every source is switched off** | Broken | A rule or dashboard reads only from objects that are disabled, so it can no longer produce anything. | Enable one of its sources, or disable it as well. |
| **The edge rejected this object** | Broken | A deployed object failed to apply on the edge gateway. | Open the object, read the edge's error, fix it and deploy again. |
| **Device has never sent data** | Needs a look | No data has arrived from an enabled device since the hub started. | Check its connection settings and that its gateway can reach it, or disable it until it is installed. |
| **Rule reads nothing we can resolve** | Needs a look | No tag, telemetry key or device matches the rule's condition. | Check that the tags or devices in the condition still exist and are spelled as configured. |
| **Pipeline is published but runs nowhere** | Needs a look | The pipeline is not assigned to a gateway, or no gateway has received it. | Assign it to a gateway and publish it again, or disable it. |
| **Pipeline reads no device** | Needs a look | The pipeline's input matches no device on its gateways. | Check that the devices named in its input still exist and run on its gateways. |
| **The edge runs a different version** | Needs a look | What is configured here is not what the edge is running. | Deploy the object, or discard the change if it was a test. |
| **Everything runs on one gateway** | Needs a look (**Risk**) | Many rules, functions and targets depend on one gateway; a restart stops them all together. | Plan maintenance around the gateway, or move part of the load to a second gateway. |
| **Tags nobody stores or reads** | Waste | Devices poll tags whose data goes to no target or storage, and no rule, function or dashboard reads them. | Send the data to a target, use the tags, or remove them from the profile. |
| **Function output goes nowhere** | Waste | A function produces values that are not stored, sent or read. | Use the output, store it, or stop producing it. |

## Findings you accept

Some findings are deliberate, for example a test device that is never connected. Select **Don't flag again** on
the finding. It moves to **… findings not flagged** at the bottom of the list, where **Flag again** brings it
back. This is saved for everyone.

## Trace an object

The **Trace an object** tab follows one object through the setup. Type a device, gateway, rule, function, pipeline
or target in the search box and choose it. See
[See what depends on an object](https://proxus.io/docs/organize/operational-graph#trace-an-object).

## What's next

- [See what depends on an object](https://proxus.io/docs/organize/operational-graph)
- [Find what stopped](https://proxus.io/docs/monitor/find-what-stopped)
- [Troubleshoot automation](https://proxus.io/docs/troubleshoot/automation)