Docs » Get started with the Splunk Distribution of the OpenTelemetry Collector » Configure the Collector » Configure pipelines

Configure pipelines 🔗

A pipeline defines a path the data follows in the Collector starting from reception, then further processing or modification, and finally when it exits the Collector through exporters. Pipelines operate on three data types: logs, traces, and metrics.

To learn more about data in Observability Cloud, see Data types in Splunk Observability Cloud.

Define the pipeline 🔗

The pipeline is constructed during Collector startup based on the pipeline definition. See Components to understand the behavior of each component, and check Collector default configuration.

To define the pipeline, first you need to specify a data type in your pipeline configuration. All the receivers, exporters, and processors you use in a pipeline must support the particular data type, otherwise you’ll get the ErrDataTypeIsNotSupported error message when the configuration is loaded.

A pipeline can contain one or more receivers. Data from all receivers is pushed to the first processor, which performs processing on it and then pushes it to the next processor and so on until the last processor in the pipeline pushes the data to the exporters. Each exporter gets a copy of each data element. The last processor uses a data fan-out connector to fan out (distribute) the data to multiple exporters.

Example of a pipeline configuration 🔗

A pipeline configuration typically looks like this:

  # Pipelines can contain multiple subsections, one per pipeline.
    # Traces is the pipeline type.
      receivers: [otlp, jaeger, zipkin]
      processors: [memory_limiter, batch]
      exporters: [otlp, jaeger, zipkin]

This example defines a pipeline for traces, with three receivers, two processors, and three exporters. The following table describes the receivers, processors, and exporters used in this example.



Pipeline type


otlp: Receives data through gRPC or HTTP using OTLP format.

Traces, metrics, logs


jaeger: Receives trace data in Jaeger format.



zipkin: Receives spans from Zipkin (V1 and V2).



memory_limiter: Prevents out of memory situations.

Metrics, traces, logs


batch: Accepts spans, metrics, or logs and places them into batches. Batching helps better compress the data and reduce the number of outgoing connections required to transmit the data.

Metrics, traces, logs


otlp: Exports data through gRPC using OTLP format. By default, this exporter requires TLS and offers queued retry capabilities.

Traces, metrics


jaeger: Exports data through gRPC to Jaeger destinations. By default, this exporter requires TLS and offers queued retry capabilities.



zipkin: Exports data to a Zipkin server. By default, this exporter requires TLS and offers queued retry capabilities.


Processing logs 🔗

See Splunk Log Observer Connect.

Processing metrics 🔗

See Metrics pipeline management in Splunk Infrastructure Monitoring to manage metrics. See also View organization metrics for Splunk Observability Cloud for metrics generated by Splunk Observability Cloud.

Processing spans, traces, and tags 🔗

See Manage services, spans, and traces in Splunk APM and Configure tags in OpenTelemetry.

Metadata transformations 🔗

Metadata refers to the name/value pair added to telemetry data. OpenTelemetry calls this Attributes on Spans, Labels on Metrics, and Fields on Logs. See Resource SDK, Metrics API, and Trace Semantic Conventions in GitHub for additional details.

Attributes 🔗

Attributes are a list of zero or more key-value pairs. An attribute must have the following properties:

  • The attribute key, which must be a non-null and non-empty string.

  • The attribute value, which is one of these types:

    • A primitive type: string, boolean, double precision floating point (IEEE 754-1985) or signed 64-bit integer.

    • An array of primitive type values. The array must be homogeneous. That is, it must not contain values of different types. For protocols that do not natively support array values, represent those values as JSON strings.

Attribute values expressing a numerical value of zero, an empty string, or an empty array are considered meaningful and must be stored and passed on to processors or exporters.

Attribute values of null are not valid and attempting to set a null value is undefined behavior.

null values are not allowed in arrays. However, if it is impossible to make sure that no null values are accepted (for example, in languages that do not have appropriate compile-time type checking), null values within arrays MUST be preserved as-is (that is, passed on to span processors/exporters as null). If exporters do not support exporting null values, you can replace those values by 0, false, or empty strings. Changing these values is required for map and dictionary structures represented as two arrays with indices that are kept in sync (for example, two attributes header_keys and header_values, both containing an array of strings to represent a mapping header_keys[i] -> header_values[i]).

Labels 🔗

Labels are name/value pairs added to metric data points. Labels are deprecated from the OpenTelemetry specification. Use attributes instead of labels.

Fields 🔗

Fields are name/value pairs added to log records. Each record contains two kinds of fields:

  • Named top-level fields of specific type and meaning.

  • Fields stored as map<string, any>, which can contain arbitrary values of different types. The keys and values for well-known fields follow semantic conventions for key names and possible values that allow all parties that work with the field to have the same interpretation of the data.