Extending Monitoring Configuration with Custom Data Points

Prometheus clients offer the flexibility to process multiple configuration files. This allows you to add custom data points and extend the default monitoring scope. Take advantage of this feature by creating a separate file containing your tailored data points. The next example showcases how to implement this by adding a new file and publishing a single data point to the Prometheus server.

Steps

Create a Custom Metrics File: Inside your project, create a file named customMetrics.yaml (or similar) within a suitable directory (e.g., data/metrics).
Note: The PromClient searches in all folders defined in your project config file.
Define Your Custom Metric: Edit customMetrics.yaml with the following YAML snippet:
```
globalVariables:
defaultLabels:

metricObjects:
  CustomValue:
    help: Custom metric value
    type: Gauge
    labelNames:
      - type
  
metricData:
  - objectType: CustomValue
    value: !dpGet
      dpId: ExampleDP_Arg1.:_original.._value
    labels:
      type: "Custom Value"
    function: set
```
This creates a metric named CustomValue of type Gauge with a single label type. Data for this metric comes from the dpGet method, referencing ExampleDP_Arg1 data point. The set function ensures the gauge reflects the latest value. This metric inherits default labels from the parent configuration.
Configure Prometheus Client: Modify your Prometheus client progs to include both configuration files:
```
promClient.js -f metrics/defaultMetrics.yaml -f
          data/metrics/customMetrics.yaml
```
Start the Client and Monitor: Start your Prometheus client. The newly defined CustomValue metric will now appear alongside default metrics in the Prometheus server.
Note: Note that Prometheus health reporting is not a replacement for proper value archiving. The poll rate for obtaining health metrics is typically set very low, and the number of archived data points is also low. Despite these limitations, the feature has been tested and confirmed to work within the expected performance limits, supporting up to 10,000 unique configured DPEs being reported from the WinCC OA system.