# 02 OEE Dashboard

In this tutorial you build a department dashboard with aggregated OEE metrics, production status for every machine, and a department overview:

<figure><img src="https://4261006941-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FSNEuiyRRKwuqtIcaEt45%2Fuploads%2Fgit-blob-e2dab229c2e58701729e0a8013f946928130c633%2Fgrafana-oee-hero-01.png?alt=media" alt=""><figcaption></figcaption></figure>

## What you will learn

* Configure POST requests with the `productivity-metrics` endpoint
* Create hidden time variables with 10-minute alignment
* Display OEE metrics as a table with color scale
* Visualize machine activity as a state timeline
* Calculate weighted OEE aggregation for department overviews

## Prerequisites

* [Query the ENLYZE API](https://docs.enlyze.com/en/integrations/grafana/advanced-api/02-api-queries) and [Variables](https://docs.enlyze.com/en/integrations/grafana/advanced-api/03-variables) completed
* Understanding of transformations ([Tutorial 3.2](https://github.com/enlyze/enlyze-docs/blob/main/en/grafana/intermediate-features/03-transformations.md))

***

## 10-minute alignment

The `productivity-metrics` endpoint requires time ranges aligned to 10-minute intervals (minutes: 0, 10, 20, 30, 40, 50; seconds and milliseconds: 0).

When users select a time range such as "Last 7 days" in Grafana, the time values contain arbitrary seconds. Create two hidden query variables that round the time values automatically.

### Create time variables

Create a new dashboard and add two variables under **Settings** > **Variables**:

| Setting     | start                 | end                  |
| ----------- | --------------------- | -------------------- |
| Name        | `start`               | `end`                |
| Type        | Query                 | Query                |
| Data source | ENLYZE API            | ENLYZE API           |
| Hide        | Variable              | Variable             |
| Refresh     | On time range change  | On time range change |
| Source      | Inline                | Inline               |
| Data        | `{"time": ${__from}}` | `{"time": ${__to}}`  |

**Root selector (JSONata, identical for both variables):**

```
(
  $time_millis := $number($.time);
  $time_rounded := $time_millis - ($time_millis % 600000);
  $fromMillis($time_rounded)
)
```

The expression converts the timestamp to milliseconds, rounds down to the nearest 10-minute interval, and converts back to an ISO date.

<figure><img src="https://4261006941-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FSNEuiyRRKwuqtIcaEt45%2Fuploads%2Fgit-blob-e2dab229c2e58701729e0a8013f946928130c633%2Fgrafana-oee-hero-01.png?alt=media" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
For simpler dashboards you can also round to full hours: `${__from:date:YYYY-MM-DDTHH:00:00}+00:00`. The 10-minute rounding is more precise, though, and works with all ENLYZE API endpoints.
{% endhint %}

***

## OEE by machine (table)

The table shows OEE, Availability, Performance, and Quality for each machine in the department.

### Configure the POST request

Create a **Table** panel and configure the first query:

| Setting           | Value                                                                |
| ----------------- | -------------------------------------------------------------------- |
| Type              | JSON                                                                 |
| Parser            | Backend                                                              |
| Method            | POST                                                                 |
| URL               | `machines/141e0927-62b3-4e76-8398-ad82d20f397f/productivity-metrics` |
| Body type         | Raw                                                                  |
| Body content type | application/json                                                     |

**Body:**

```json
{
  "datetime_ranges": [
    {
      "start": "${start}",
      "end": "${end}"
    }
  ]
}
```

<figure><img src="https://4261006941-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FSNEuiyRRKwuqtIcaEt45%2Fuploads%2Fgit-blob-e9edc3c098dacbc693ba4187b14ed323ff32e841%2Fgrafana-oee-post-request-01.png?alt=media" alt=""><figcaption></figcaption></figure>

{% hint style="warning" %}
Make sure the **Body content type** is set to `application/json` (not `text/plain`). Otherwise the API responds with HTTP 422.
{% endhint %}

### Columns and computed columns

Set the **Root Selector** to `$.data` and add the following **Columns** under **Parsing options & Result fields**:

| Selector             | Title        | Type   |
| -------------------- | ------------ | ------ |
| `productivity.score` | productivity | Number |
| `availability.score` | availability | Number |
| `performance.score`  | performance  | Number |
| `quality.score`      | quality      | Number |

Add two **Computed Columns** to identify the machine:

| Selector                                 | Title         | Type   |
| ---------------------------------------- | ------------- | ------ |
| `"141e0927-62b3-4e76-8398-ad82d20f397f"` | machine\_uuid | String |
| `"Kiefel"`                               | Machine       | String |

### Additional machines

Repeat the query for each machine in the department (Macchi, Alpine, Reifenhäuser, W\&H Varex). Change the UUID in the URL and in the computed columns accordingly.

### Transformations

1. **Merge** combines all queries into a single table
2. **Organize fields** hides `machine_uuid` and renames the columns (productivity to OEE, availability to Availability, etc.)

### Visual styling

Configure the table under **Standard options**:

* **Unit**: Percent (0.0-1.0)
* **Color scheme**: Continuous - RdYlGr
* **Min**: 0, **Max**: 1

Create **Field Overrides** for the OEE, Availability, Performance, and Quality columns:

* **Cell type**: Gauge (basic, color display)

Enable the **Table Footer** with Calculation **Mean** for the department averages.

Add a **Value mapping**:

* **Type**: Special > Null
* **Display text**: NO DATA
* **Color**: Text

<figure><img src="https://4261006941-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FSNEuiyRRKwuqtIcaEt45%2Fuploads%2Fgit-blob-fcc7e2b577f1a8850cc8495494d7fa345ed449c7%2Fgrafana-oee-table-01.png?alt=media" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
Not all machines provide quality data. In the demo account, Quality is 1 (100%) for all machines. The `null` mapping ensures that machines without data in the selected time range show "NO DATA" instead of an error.
{% endhint %}

***

## Machine activity (state timeline)

The state timeline shows at a glance when each machine was in production and when it was idle.

### Use the ENLYZE data source

Create a **State Timeline** panel with the **ENLYZE** data source (not ENLYZE API). Create a query for each machine and select the **Durchsatz** (throughput) variable.

Configure **Field Overrides** for each query (by **Frame refID**) to set the Display Name to the machine name (e.g. "Kiefel", "Macchi", etc.).

### Value mappings

Create two range mappings to convert the throughput value into an operating status:

| From  | To    | Text      | Color |
| ----- | ----- | --------- | ----- |
| -1000 | 1     | IDLE      | Red   |
| 1     | 10000 | PRODUCING | Green |

The logic: if throughput is at or near 0, the machine is idle. If it is above 1, it is producing.

<figure><img src="https://4261006941-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FSNEuiyRRKwuqtIcaEt45%2Fuploads%2Fgit-blob-05c8cb683cb24a7c10cc45aec79acc559d880c74%2Fgrafana-oee-state-timeline-01.png?alt=media" alt=""><figcaption></figcaption></figure>

***

## Department overview

Four stat panels show the aggregated OEE metrics for the entire department: OEE, Availability, Performance, and Quality.

### OEE and Availability (mean)

Create a **Stat** panel for the Department OEE. Configure 5 queries (one per machine) with the same POST requests as in the table, but with a different root selector and columns:

* **Root Selector**: `$.data`
* **Column**: `productivity.score` as "OEE" (Number)

Add a **Merge** transformation. Under **Value options** > **Calculation**, select **Mean**.

Configure:

* **Color mode**: Background
* **Unit**: Percent (0.0-1.0)
* **Thresholds**: Red (base), Yellow (0.6), Green (0.8)
* **Value Mapping**: `null` to "NO DATA"

Create a second panel for **Availability** following the same pattern, with Column `availability.score`.

### Performance and Quality (weighted)

For Performance and Quality, a simple mean is not sufficient. When machines run for different amounts of time, the values must be weighted by Availability.

**Formula:**

$$
\text{Dept. Performance} = \frac{\sum\_{i} \text{Availability}\_i \times \text{Performance}*i}{\sum*{i} \text{Availability}\_i}
$$

Create a **Stat** panel for Performance. The 5 queries use a JSONata root selector:

```
$map($.data, function($v) {
  {
    "avail": $v.availability.score,
    "weighted_perf": $v.availability.score * $v.performance.score
  }
})
```

Each query returns the Availability and the weighted product (Availability x Performance).

**Transformations:**

1. **Merge** combines all queries
2. **Reduce** with Mode **Reduce fields** and Calculation **Sum** sums both columns
3. **Add field from calculation** with Mode **Binary operation**: `weighted_perf / avail`, Alias "Performance", **Replace all fields** enabled

For **Quality**, follow the same pattern with this root selector:

```
$map($.data, function($v) {
  {
    "avail": $v.availability.score,
    "weighted_quality": $v.availability.score *
      ($exists($v.quality) ? $v.quality.score : 1)
  }
})
```

The `$exists` fallback ensures that machines without quality data are treated as Quality = 1 (100%).

<figure><img src="https://4261006941-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FSNEuiyRRKwuqtIcaEt45%2Fuploads%2Fgit-blob-0162b03e04e816e1ca810233d35a2d6a70137177%2Fgrafana-oee-dept-summary-01.png?alt=media" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}
The weighted aggregation ensures that machines with high availability contribute more to the department metrics than machines that barely ran. For more on OEE calculation, see the [ENLYZE documentation](https://docs.enlyze.com).
{% endhint %}

***

## Tips

* **10-minute variables**: The hidden `start` and `end` variables work in any dashboard that uses the `productivity-metrics` endpoint. Copy them to other dashboards as needed.
* **production-runs vs. productivity-metrics**: Use `production-runs` (GET, [Tutorial 4.2](https://docs.enlyze.com/en/integrations/grafana/advanced-api/02-api-queries)) for OEE per production run. Use `productivity-metrics` (POST) for aggregated shift or daily reports.
* **422 Unprocessable Entity**: The time values are not correctly aligned. Check the hidden variables and make sure the body content type is set to `application/json`.
* **Single machine**: For a dashboard with the live status of a single machine, see [Production status dashboard](https://docs.enlyze.com/en/integrations/grafana/production-dashboards/04-production-status).

***

## Next steps

* [**Downtime reporting**](https://docs.enlyze.com/en/integrations/grafana/production-dashboards/06-downtime-reporting) -- Analyze and visualize machine downtimes