acutracer (Distributed tracer)

A llm flow tracer

The following repo will have the instrumentors for a agentic system.

The instrumentor will trace all the outgoing calls made through a agentic system.

The following headers will be added to the calls.

X-acuvity-trace-id X-acuvity-span-id X-acuvity-parent-id

The headers will be captured by the proxy for further processing.

Running pytest

From the root directory run the following

pytest

Visualization of the traces

All the traces are OTEL compatible, so any OTEL collector can be used to view the traces locally.

Supported Instrumentors for python

Web Frameworks:
- FastAPIInstrumentor
- FlaskInstrumentor
- DjangoInstrumentor
HTTP Clients:
- RequestsInstrumentor
- URLLibInstrumentor
- AioHTTPClientInstrumentor

Future work for Granular Instrumentors in python

Databases:
- MongoInstrumentor
gRPC:
- GRPCInstrumentor
AWS:
- BotoInstrumentor
Logging:
- LoggingInstrumentor

Jaeger example with FAST API instrumentor:

Setting up the app:

1st pip install the acutracer

pip install acutracer

Import the following lines.

from acutracer.instrumentors.python.webapi.instrumentor import (
    WebAPIInstrumentor,
)

app = FastAPI(title="Fast-API")

instrumentor = WebAPIInstrumentor(name="acuvity-tracing-example")
tracer = instrumentor.instrument(app=app)

Setting Up and Using Jaeger UI (for macOS)

Starting Jaeger

Run the following command to start the Jaeger all-in-one Docker container:

docker run -d --name jaeger \
  -e COLLECTOR_ZIPKIN_HOST_PORT=:9411 \
  -e COLLECTOR_OTLP_ENABLED=true \
  -p 6831:6831/udp \
  -p 6832:6832/udp \
  -p 5778:5778 \
  -p 16686:16686 \
  -p 14317:4317 \
  -p 4318:4318 \
  -p 14250:14250 \
  -p 14268:14268 \
  -p 14269:14269 \
  -p 9411:9411 \
  jaegertracing/all-in-one:latest

Note: We've changed -p 4317:4317 to -p 14317:4317. This maps the container's internal port 4317 to the host's port 14317.

Accessing the Jaeger UI

Once the container is running, you can access the Jaeger UI by opening a web browser and navigating to:

http://localhost:16686

Viewing Traces

Start your instrumented Python application.
Make some requests to your application to generate traces.
Open the Jaeger UI in your web browser.
In the UI:
- Select your service from the "Service" dropdown (it should be named after your Python application).
- Set an appropriate time range.
- Click "Find Traces" to view your traces.

Troubleshooting

If you don't see any traces:

Check that your application is running and generating traffic.
Ensure that the Jaeger agent port (6831) is accessible from your application.
Verify that the exporter in your Python code is correctly configured to send traces to localhost:6831.
Check the console output of your Python application for any error messages related to trace export.

Remember to stop the Jaeger container when you're done:

docker stop jaeger
docker rm jaeger

Metadata

Release files for acutracer 0.1.29

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for acutracer 0.1.29
File	Size	Uploaded
acutracer-0.1.29.tar.gz	19.8 kB	Sep 16, 2024	Details

Built distribution (wheel)

Table of built distributions (wheels) for acutracer 0.1.29
File	Interpreter	ABI	Platform	Reset
acutracer-0.1.29-py3-none-any.whl 20.8 kB Sep 16, 2024	Python 3	none	any	Details

Total release size: 40.6 kB

Release files / acutracer-0.1.29.tar.gz

Download URL	acutracer-0.1.29.tar.gz
Size	19.8 kB
Tags	Source
SHA-256 checksum How to use checksums	`3247f19033994db1fc3e831c74c5c2811c32d409cced1ea731cc488c2bb2cb2f`
BLAKE2b-256 checksum How to use checksums	`3d6971f755b6635ca6cc8719e956c608ec16aab30cf27fff335d170d317826c6`
Upload date	Sep 16, 2024
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/5.1.1 CPython/3.11.9`

Release files / acutracer-0.1.29-py3-none-any.whl

Download URL	acutracer-0.1.29-py3-none-any.whl
Size	20.8 kB
Tags	Python 3
SHA-256 checksum How to use checksums	`067ee6be4bb9ab4ed565581bd50310c68fa230e9a89997a49220202f9b8d07cb`
BLAKE2b-256 checksum How to use checksums	`f08157aa3995a0dc8c6e1ac76362d154a43b49373588a34f32b99a5b4a630765`
Upload date	Sep 16, 2024
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/5.1.1 CPython/3.11.9`