Skip to content
This repository has been archived by the owner on Jul 11, 2022. It is now read-only.


Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

Latest commit

Signed-off-by: Yuri Shkuro <>

Git stats


Failed to load latest commit information.
Latest commit message
Commit time

Build Status Coverage Status PyPI Version Python Version FOSSA Status

πŸ›‘ This library is DEPRECATED!

There will be no new releases of this library.

We urge all users to migrate to OpenTelemetry. Please refer to the notice in the documentation for details.

Jaeger Bindings for Python OpenTracing API

This is a client-side library that can be used to instrument Python apps for distributed trace collection, and to send those traces to Jaeger. See the OpenTracing Python API for additional detail.

Contributing and Developing

Please see


pip install jaeger-client

Getting Started

import logging
import time
from jaeger_client import Config

if __name__ == "__main__":
    log_level = logging.DEBUG
    logging.getLogger('').handlers = []
    logging.basicConfig(format='%(asctime)s %(message)s', level=log_level)

    config = Config(
        config={ # usually read from some yaml config
            'sampler': {
                'type': 'const',
                'param': 1,
            'logging': True,
    # this call also sets opentracing.tracer
    tracer = config.initialize_tracer()

    with tracer.start_span('TestSpan') as span:
        span.log_kv({'event': 'test message', 'life': 42})

        with tracer.start_span('ChildSpan', child_of=span) as child_span:
            child_span.log_kv({'event': 'down below'})

    time.sleep(2)   # yield to IOLoop to flush the spans -
    tracer.close()  # flush any buffered spans

NOTE: If you're using the Jaeger all-in-one Docker image (or similar) and want to run Jaeger in a separate container from your app, use the code below to define the host and port that the Jaeger agent is running on. Note that this is not recommended, as Jaeger sends spans over UDP and UDP does not guarantee delivery. (See this thread for more details.)

    config = Config(
        config={ # usually read from some yaml config
            'sampler': {
                'type': 'const',
                'param': 1,
            'local_agent': {
                'reporting_host': 'your-reporting-host',
                'reporting_port': 'your-reporting-port',
            'logging': True,

Other Instrumentation

The OpenTracing Registry has many modules that provide explicit instrumentation support for popular frameworks like Django and Flask.

At Uber we are mostly using the opentracing_instrumentation module that provides:

  • explicit instrumentation for HTTP servers, and
  • implicit (monkey-patched) instrumentation for several popular libraries like urllib2, redis, requests, some SQL clients, etc.

Initialization & Configuration

Note: do not initialize the tracer during import, it may cause a deadlock (see issues #31, #60). Instead define a function that returns a tracer (see example below) and call that function explicitly after all the imports are done.

Also note that using gevent.monkey in asyncio-based applications (python 3+) may need to pass current event loop explicitly (see issue #256):

from tornado import ioloop
from jaeger_client import Config

config = Config(config={}, service_name='your-app-name', validate=True)


The recommended way to initialize the tracer for production use:

from jaeger_client import Config

def init_jaeger_tracer(service_name='your-app-name'):
    config = Config(config={}, service_name=service_name, validate=True)
    return config.initialize_tracer()

Note that the call initialize_tracer() also sets the opentracing.tracer global variable. If you need to create additional tracers (e.g., to create spans on the client side for remote services that are not instrumented), use the new_tracer() method.

Prometheus metrics

This module brings a Prometheus integration to the internal Jaeger metrics. The way to initialize the tracer with Prometheus metrics:

from jaeger_client.metrics.prometheus import PrometheusMetricsFactory

config = Config(
tracer = config.initialize_tracer()

Note that the optional argument service_name_label to the factory constructor will force it to tag all Jaeger client metrics with a label service: your-app-name. This way you can distinguish Jaeger client metrics produced by different services.


For development, some parameters can be passed via config dictionary, as in the Getting Started example above. For more details please see the Config class.

WSGI, multi-processing, fork(2)

When using this library in applications that fork child processes to handle individual requests, such as with WSGI / PEP 3333, care must be taken when initializing the tracer. When Jaeger tracer is initialized, it may start a new background thread. If the process later forks, it might cause issues or hang the application (due to exclusive lock on the interpreter). Therefore, it is recommended that the tracer is not initialized until after the child processes are forked. Depending on the WSGI framework you might be able to use @postfork decorator to delay tracer initialization (see also issues #31, #60).

Debug Traces (Forced Sampling)


The OpenTracing API defines a sampling.priority standard tag that can be used to affect the sampling of a span and its children:

from opentracing.ext import tags as ext_tags

span.set_tag(ext_tags.SAMPLING_PRIORITY, 1)

Via HTTP Headers

Jaeger Tracer also understands a special HTTP Header jaeger-debug-id, which can be set in the incoming request, e.g.

curl -H "jaeger-debug-id: some-correlation-id"

When Jaeger sees this header in the request that otherwise has no tracing context, it ensures that the new trace started for this request will be sampled in the "debug" mode (meaning it should survive all downsampling that might happen in the collection pipeline), and the root span will have a tag as if this statement was executed:

span.set_tag('jaeger-debug-id', 'some-correlation-id')

This allows using Jaeger UI to find the trace by this tag.

Zipkin Compatibility

To use this library directly with other Zipkin libraries & backend, you can provide the configuration property propagation: 'b3' and the X-B3-* HTTP headers will be supported.

The B3 codec assumes it will receive lowercase HTTP headers, as this seems to be the standard in the popular frameworks like Flask and Django. Please make sure your framework does the same.


Apache 2.0 License.