Get started

Follow these steps to configure and activate the component:

1. Deploy the Splunk Distribution of the OpenTelemetry Collector to your Windows host platform as described in Install the Collector for Windows with the installer script.

2. Configure the Windows Performance Counters receiver as described in the next section.

3. Restart the Collector.

Sample configuration

To activate the Windows Performance Counters receiver add a windowsperfcounters entry inside the receivers section of the Collector configuration file. For example:

receivers:
  windowsperfcounters:
    metrics:
      bytes.committed:
        description: the number of bytes committed to memory
        unit: By
        gauge:
    collection_interval: 30s
    perfcounters:
    - object: Memory
      counters:
        - name: Committed Bytes
          metric: bytes.committed

To complete the configuration, include the receiver in the metrics pipeline of the service section of your configuration file:

service:
  pipelines:
    metrics:
      receivers:
        - windowsperfcounters

To collect metrics from Windows performance counters, you need to define metrics using the metrics field as in the example. You can then reference the metrics you defined from the counters.metric fields.

View available performance counters

To see a list of available performance counters, use Windows PowerShell or the Windows Performance Monitor.

PowerShell

In PowerShell, run the following command to list all performance counter sets:

Get-Counter -ListSet *

To list the instances of each performance counter set, run the following command and replace <perf_object_name> with the name of the instances you want to find:

Get-Counter -List "<perf_object_name>"

Windows Performance Monitor

Run the following command to open the Windows Performance Monitor:

perfmon /sys

In the Windows Performance Monitor, select the green plus arrow to see a list of available performance counters.

Note the following:

• If a specific performance counter can’t be accessed at startup, the receiver emits a warning and continues execution.

• Some performance counters might not exist on some systems due to different OS configuration.

Advanced configuration

Configure collection interval and counters

You can configure the collection interval and which performance counters you want to scrape. For example:

windowsperfcounters:
  collection_interval: <duration>
  initial_delay: <duration>
  metrics:
    <metric name 1>:
      description: <description>
      unit: <unit type>
      gauge: null
    <metric name 2>:
      description: <description>
      unit: <unit type>
      sum: null
      aggregation: <cumulative or delta>
      monotonic: <true or false>
  perfcounters:
  - object: <object name>
    instances:
      - <instance name>
    counters:
      - name: <counter name>
        metric: <metric name>
        attributes:
        <key>: <value>

Scrape at different collection intervals

The following example shows how to scrape performance counters using different collection intervals depending on the target:

receivers:
  windowsperfcounters/memory:
    metrics:
      bytes.committed:
        description: Number of bytes committed to memory
        unit: By
        gauge:
    collection_interval: 30s
    perfcounters:
      - object: Memory
        counters:
          - name: Committed Bytes
            metric: bytes.committed

  windowsperfcounters/processor:
    collection_interval: 1m
    metrics:
      processor.time:
        description: CPU active and idle time
        unit: "%"
        gauge:
    perfcounters:
      - object: "Processor"
        instances: "*"
        counters:
          - name: "% Processor Time"
            metric: processor.time
            attributes:
              state: active
      - object: "Processor"
        instances: ["0", "1"]
        counters:
          - name: "% Idle Time"
            metric: processor.time
            attributes:
              state: idle

  # ...

service:
  pipelines:
    metrics:
      receivers: [windowsperfcounters/memory, windowsperfcounters/processor]

Configure instances

An instance is any entity that produces performance data. Instances can have one or more counter values.

The receiver supports the following values through the instances field:

Total instance behavior and the aggregation counter

To avoid dropping the _Total instance, configure the receiver to collect it individually on its own metric, as in the following example:

windowsperfcounters:
  metrics:
    processor.time.total:
      description: Total CPU active and idle time
      unit: "%"
      gauge:
  collection_interval: 30s
  perfcounters:
    - object: "Processor"
      instances:
        - "_Total"
      counters:
        - name: "% Processor Time"
          metric: processor.time.total

Recreate the query on every scrape

On some versions of Windows, Counters are sometimes corrupted and continuously return invalid data after the first scrape. If this happens, set the counter setting recreate_query to true (defaults to false) to tell the receiver to recreate the PDH query on every scrape. This might affect performance but should be inconsequential unless collection_interval is very high.

If recreating the query fails, the previous query is reused and an error is logged.

Define metric format

Configured metrics consist of a metric description, including unit and type, used by one or more performance counters scraped by the receiver.

To report metrics in a specific format, define the metric and reference it in the corresponding counter, along with any applicable attributes. By default, the metric name corresponds to the name of the counter.

Metrics can be of type sum or gauge. Sum metrics support the aggregation and monotonic fields.

Sum metrics

The following settings apply to sum metrics:

Gauge metrics

The gauge configuration doesn’t accept settings. It is specified as an object for forwards compatibility.

The following example emits the Memory/Committed Bytes counter as the bytes.committed metric:

receivers:
  windowsperfcounters:
    metrics:
      bytes.committed:
        description: the number of bytes committed to memory
        unit: By
        gauge:
    collection_interval: 30s
    perfcounters:
    - object: Memory
      counters:
        - name: Committed Bytes
          metric: bytes.committed

service:
  pipelines:
    metrics:
      receivers: [windowsperfcounters]

Known limitations

• The network interface is not available inside the container and as a result metrics for the object Network Interface aren’t generated in that scenario. If there is a sub-process it captures Network Interface metrics.

• The counter category Process is unreliable with multiple instances of the same process.

  • For Windows 11 and higher, use Process V2 instead as it includes the process ID in the instance name.

  • For versions prior to Windows 11, you can configure the Process counter category to include the PID in the instance name. Learn more at Microsoft’s doc Handling Duplicate Instance Names .

Settings

The following table shows the configuration options for the Windows Performance Counters receiver:

Troubleshooting