> ## Documentation Index
> Fetch the complete documentation index at: https://docs.spojit.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Create a monitor

> Draft a monitor with Miraxa, start from a recommended one, or build it in the one-sentence editor.

You need permission to manage monitors (Owners and Admins have it by default). Start from the **Monitors** tab, or **New monitor** on the Incidents tab.

## Three ways to start

**1. Draft it with Miraxa**

Type what you want watched in your own words, such as *"Make sure my orders are going through"* or *"Alert me if any workflow fails 3 times in a row"*, and choose **Draft it with Miraxa**. Miraxa drafts the monitor in a chat and shows it as a card. Choose **Create monitor** to save it, **Edit first** to open it in the editor, or **Cancel**.

**2. Start from a recommended monitor**

**Recommended for this workspace** lists ready-made monitors, such as workflows that stop succeeding, stuck runs, a high error rate, low credits, and AI credits or execution minutes nearing your allowance. **Set up** opens one in the editor so you can adjust it before saving.

**3. Build it yourself**

Give it a **Name**, choose what it should do when it finds something, then set each piece of the sentence.

## Tell me once, or alert me while it's wrong

The editor first asks **What should it do when it finds something?**

* **Tell me once** raises a [notice](https://docs.spojit.com/control-center/overview.md#incidents-and-notices) each time something new appears: a new API version, a price change, a broadcast.
* **Alert me while it's wrong** raises an incident that stays open until it clears: failing runs, low credits, a value over a limit. Reminders follow if it drags on.

The choice decides which [sources](https://docs.spojit.com/control-center/sources.md) you can pick. Changing it on a saved monitor changes what the monitor watches, so it starts again from its next check.

## The monitor, in one sentence

The editor shows the whole monitor as a sentence, for example:

> Open an incident when **workflows stop succeeding** in **all active workflows**, checking **every 15 minutes**, and tell **Ops in Slack** as **Warning**.

Select any highlighted piece to change it:

| Piece | What you set |
| - | - |
| **What it watches** | The [source](https://docs.spojit.com/control-center/sources.md) and its settings, such as which workflows, which page or which connection. |
| **Checking** | How often it checks: every so often, or at [set times](https://docs.spojit.com/control-center/schedules.md). |
| **Who to tell** | Email me, your [contacts](https://docs.spojit.com/control-center/alerts.md), and whether to also tell them when a fix is applied. |
| **How serious** | Critical, Warning or Info. Notices default to Info; usage, web page and connection incidents to Warning; other incidents to Critical. |

## More options

| Option | What it does |
| - | - |
| **Incident timing** | For incident monitors: how long a problem must hold before an incident opens, how long it stays open after it clears, what to do when there is no data, and how often to send reminders. |
| **Suggested next steps** | Ask Miraxa to look into each alert and suggest a fix. Uses AI credits. See [Suggested fixes](https://docs.spojit.com/control-center/fixes.md). |
| **Auto-apply suggested fixes** | Let low-risk (or low and medium risk) fixes apply on their own. |
| **Ask Miraxa limits** | For monitors that ask Miraxa: the web pages it may read, whether it may search the web, the time zone, and how many credits a check or a day may use. |
| **Start paused** | Save the monitor without it checking yet. On a saved monitor this is the **Active** switch. |
| **Description** | A note for your team about why the monitor exists. |

### Incident timing

| Setting | What it does |
| - | - |
| **Only open it if it holds for** | Waits for the problem to last this long before opening an incident, so one blip does not alert. |
| **Keep it open after it clears for** | Keeps the incident open this long after things look normal, so a flapping problem does not open and close repeatedly. |
| **When there is no data** | Show **No data**, treat it as all clear, treat it as wrong, or keep the last state. |
| **Remind every** | How often to remind people while the incident stays open. One hour by default. |

## Test it now

Before saving, **Test it now** runs one check exactly as it will run when saved, and saves nothing. For a monitor that asks Miraxa, **Test check** shows whether it **would alert**, **would stay quiet**, or **could not tell**, and what the check cost. Asking Miraxa uses AI credits.

## Pausing and editing

The **Active** switch on the Monitors tab pauses or resumes a monitor. Open a monitor to see its page: the sentence, how often it checks, its last check, the AI credits it has used, and its history. Incident monitors also show a 7-day graph and what each workflow or value looks like now. **Subscribe me** adds you to its alerts without editing it.

If a monitor's own checks keep failing (for example, a watched page is gone), it opens a *can't check* incident at its own level, which resolves at its next good check or when you save the monitor.

> [card] [Try this on the free plan](https://miraxa.spojit.com/signup)
>
> 500 AI credits and 200 workflow runs a month, no card needed.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.