OpenTelemetry Baggage Span Processor

The BaggageSpanProcessor reads entries stored in Baggage from the parent context and adds the baggage entries' keys and values to the span as attributes on span start.

Installation

pip install opentelemetry-processor-baggage

Add this span processor to a tracer provider.

Keys and values added to Baggage will appear on subsequent child spans for a trace within this service and be propagated to external services in accordance with any configured propagation formats configured. If the external services also have a Baggage span processor, the keys and values will appear in those child spans as well.

[!WARNING]

Do not put sensitive information in Baggage.

To repeat: a consequence of adding data to Baggage is that the keys and values will appear in all outgoing HTTP headers from the application.

## Usage

Add the span processor when configuring the tracer provider.

To configure the span processor to copy all baggage entries during configuration:

from opentelemetry.processor.baggage import BaggageSpanProcessor, ALLOW_ALL_BAGGAGE_KEYS

tracer_provider = TracerProvider()
tracer_provider.add_span_processor(BaggageSpanProcessor(ALLOW_ALL_BAGGAGE_KEYS))

Alternatively, you can provide a custom baggage key predicate to select which baggage keys you want to copy.

For example, to only copy baggage entries that start with my-key:

starts_with_predicate = lambda baggage_key: baggage_key.startswith("my-key")
tracer_provider.add_span_processor(BaggageSpanProcessor(starts_with_predicate))

For example, to only copy baggage entries that match the regex ^key.+:

regex_predicate = lambda baggage_key: baggage_key.startswith("^key.+")
tracer_provider.add_span_processor(BaggageSpanProcessor(regex_predicate))

References

• OpenTelemetry Project