page_type	languages	products	name	description	urlFragment
sample	python	azure	Use OpenCensus with Python FAST API and Azure Monitor	Using OpenCensus with Python FastAPI and Azure Monitor for logging and tracing	opencensus-with-fastapi-and-azure-monitor

Using OpenCensus with Python FAST API and Azure Monitor

Azure Monitor supports distributed tracing, metric collection, and logging of Python applications through integration with OpenCensus.

This sample application walks you through the process of setting up OpenCensus for Python FAST API and sending monitoring data to Azure Monitor for logging and tracing.

Tracing

• Hitting an endpoint in the FastAPI application will be treated as an incoming requests (requests table in Azure Monitor).

Logging

• A log message is also sent every time a function is called (traces table in Azure Monitor).
• An exception telemetry is sent when an invalid task is entered (established date equals to 0).

Metrics (partially supported via OpenCensus.stats)

• Custom metrics are implemented on a specific function in the sample (customMetrics table in Azure Monitor).

NOTE: OpenCensus.stats supports 4 aggregation methods but provides partial support for Azure Monitor. Distribution aggregation for example, is NOT supported by the Azure Exporter.

NOTE: log_level sets the threshold for logger to "selected level", In this sample logger level is set to info. So all traces will reflecting into Azure Monitor.

if __name__=="__main__":
    print("main started")
    uvicorn.run("main:app", port=8000, log_level="info")

You can reach out to all options of uvicorn in Uvicorn deployment documentation

Pre-requisites

• Python 3.x
• Microsoft Azure Account

1. Running Clubs API project

1. Create a virtual environment.

python -m venv .venv

2. Activate the virtual environment you just created.

source .venv/bin/activate

3. Rename the .env.sample.txt file to .env
   This contains the environment variables that you need to set for your Azure Storage account.

4. Set Application Insights Instrumentation Key variables in the .env file:
   APPINSIGHTS_INSTRUMENTATIONKEY

5. Install the dependencies.

pip install -r requirements.txt

6. Run the project!

python src/main.py

Visit http://127.0.0.1:8000/docs for Swagger UI

main.py Swagger UI:

1. GET /

curl -X 'GET'   'http://127.0.0.1:8000'   -H 'accept: application/json'   -H 'Content-Type: application/json'

result:

"Clubs API"

Create Telemetry data

1. Create a new club POST /clubs

curl -X 'POST' \
  'http://127.0.0.1:8000/clubs' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
  "name": "Fenerbahce",
  "country": "Turkey",
  "established": 1907
}'

curl -X 'POST' \
  'http://127.0.0.1:8000/clubs' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
  "name": "Manchester United",
  "country": "United Kingdom",
  "established": 1878
}'

The following code in create_club function creates a custom property in Azure Monitor.

## CustomDimentions
properties = {'custom_dimensions': club_db[-1] }
logger.warning('club record is added', extra=properties)

Custom dimension result will look like below:

2. Create an exception with wrong type: POST /clubs

curl -X 'POST' \
  'http://127.0.0.1:8000/clubs' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
  "name": "Arsenal",
  "country": "United Kingdom",
  "established": "a"
}'

result will be:

{"detail":[{"loc":["body","established"],"msg":"value is not a valid integer","type":"type_error.integer"}]}

3. Create an exception with invalid establish date: POST /clubs

curl -X 'POST' \
  'http://127.0.0.1:8000/clubs' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
  "name": "Arsenal",
  "country": "United Kingdom",
  "established": 0
}'

result will be:

"Internal Server Error"

4. GET /clubs

curl -X 'GET'   'http://127.0.0.1:8000/clubs'   -H 'accept: application/json'   -H 'Content-Type: application/json'

5. DELETE /clubs

curl -X 'DELETE' \
  'http://127.0.0.1:8000/clubs/Fenerbahce' \
  -H 'accept: application/json'

6. Let's create another exception with DELETE /clubs

curl -X 'DELETE' \
  'http://127.0.0.1:8000/clubs/Chelsea' \
  -H 'accept: application/json'

Since this record is not available we get a result like below:

"An error accured while deleting club id:[], Chelsea"

7. In this scenario, in addition to Logs and Traces, let's generate some custom metrics and push into Azure Monitor using OpenCensus stats. Stats are partially supported in Azure Monitor for more information visit OpenCensus Python Metrics - Microsoft Docs

GET /log_custom_metric

curl -X 'GET'   'http://127.0.0.1:8000/log_custom_metric'   -H 'accept: application/json'   -H 'Content-Type: application/json'

After genering these custom metrics, we can view these in customMetrics in Azure Monitor.

Results will look like below on Azure Monitor:

3. Running Tests

pytest tests/test.py

$ pytest tests/test.py
=========================== test session starts =================================
platform win32 -- Python 3.8.7, pytest-6.2.2, py-1.10.0, pluggy-0.13.1
rootdir: C:\Users\username\opencensus-with-fastapi-and-azure-monitor
collected 7 items

tests\test.py ......                                                       [100%]
================================ 7 passed in 0.48s ===============================

License

See LICENSE.

Code of Conduct

This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact opencode@microsoft.com with any additional questions or comments.

Contributing

See CONTRIBUTING.