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

# Test routines

> Build the checks Room Readiness runs to prove a room works.

A test routine is a list of steps that a readiness check runs in a room at its set time each morning: power devices on, wait for them to boot, check they're online, check a reading such as a microphone level, send a command such as dialling a test call. If every step passes, the room is ready.

One routine can serve many rooms. Set the default on the **Readiness checks** tab of [Room Readiness](https://docs.marcus.involvecloud.com/portal/room-readiness.md#readiness-checks), and give particular rooms their own with **Customise**.

You need **Manage Room Readiness** to create or change routines.

## Create a routine

1. In **Room Readiness**, open the **Readiness checks** tab and select **Manage routines**.
2. Select **Blank routine**, or **From standard template** for a worked example that powers the room on, dials a test call from the codec, checks the microphones hear the returning audio, then hangs up. Edit its SIP address, and the DSP reading and preset command to match the names set up on your DSP.
3. Give it a **Name** and a one-line **Description**.
4. Add steps from the list on the right, then fill in each one. Drag steps to reorder them.
5. Select **Save changes**.

The room picker above the step list is only a preview: it greys out step types that no device in that room supports, so you can see what a room can do. A routine isn't tied to a room.

## Steps

| Step | What it does | Passes when |
| - | - | - |
| **Section** | A heading to group the steps that follow. | Always |
| **Wait** | Pauses for a number of seconds, for example while a codec boots. | Always |
| **Power on** / **Power off** | Sends power on or off to the target devices that support it; others, such as PDUs, are skipped. | Every device accepts the command |
| **Device command** | Sends a named command, such as `dial` or `recall_preset`, with optional parameters. The commands offered are the ones the target devices support. | Every device completes the command |
| **Expect status** | Checks the target devices are online, or offline. | Every device has that status |
| **Check metric** | Compares a reading, such as a microphone input level, against a threshold: greater than, at least, less than, at most or equal to. | Every device's reading meets the threshold |
| **Ping** | Pings each target device's address from its Collector, whatever kind of device it is. You can set a maximum packet loss and response time. | Every device replies, within any limits set |

### Ping

**Ping** checks the network path to a device, not the device's own control connection. Use it on one device or on the whole room. Devices on a serial port have no address to ping, so a whole-room or device-type ping leaves them out.

Each device gets 3 pings by default, up to 10, and each ping has a second to reply. With no limits set, any reply passes. The result shows each device's replies and average response time, such as **Replied 3/3, 4 ms**.

Ping steps need Collector version 1.0.69 or later. On an older Collector the step fails and says the Collector needs updating.

### Up-to-date readings

**Expect status** and **Check metric** ask the Collector to check each device there and then, so they see the room as it is now, not as it was at the last scheduled check, which can be several minutes earlier. Allow up to a minute for a step that checks many devices.

PDUs are the exception. They aren't checked on demand, because they log in for each reading; steps use their latest regular reading instead, which can be up to 10 minutes old.

If a device's Collector can't take a reading (it's offline, or older than version 1.0.35), the step uses the device's last report instead and says so if it fails. For **Check metric** that last report must be from within the **Sample window**: 5 seconds for a new step, or 30 seconds if you leave it empty. Update older Collectors from the **Collectors** page.

### Choose the devices

Every step except **Section** and **Wait** has a **Target**:

* **Whole room**: every device in the room;
* **By device type**: every display, camera, video codec, DSP / audio, control panel, AV-over-IP or power / PDU device in the room;
* **Specific device**: one named device.

Steps only ever reach devices in the room being tested. Targeting by type lets one routine work across rooms with different equipment.

### When a step fails

Set **On failure** on each step:

| Option | Effect |
| - | - |
| **abort** | Stops the routine. Use it where later steps depend on this one, such as powering on. |
| **continue** | Records the failure and carries on, so one run shows every problem. |
| **retry:1**, **retry:3** | Tries the step again up to that many times before failing it. |

Any failed step marks the room as **Failed**, whichever option you choose. **Power on**, **Power off** and **Device command** steps have a **Timeout (seconds)**; one that runs past it fails. **Check metric** has a **Sample window (seconds)** instead.

### Diagram

Select **Diagram** to see the routine as a flow chart: each step in order, sections as bands around their steps, and what happens when each step fails (the run stops, carries on, or retries). It shows the routine as it is in the builder, including changes you haven't saved yet. Download it as **PNG** for documents and email, or **SVG** to scale it in Word or PowerPoint.

### Advanced JSON

Select **Advanced JSON** to edit the steps as JSON, for example to copy a routine between environments. Switch back to **Builder** to carry on visually.

## Try it on a room

Select **Test on a room**, pick a room and start. The routine runs straight away against that room, whatever routine it normally uses, and you're taken to the run to watch each step's result arrive. Save your changes first.

Updated 06 Nov 2026 14:07 · page 459294b


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