Fix issue with Flask instrumentation when a request spawn children threads and copies the request context

[hangonlyra]

Description

In opentelemetry.instrumentation.flask record the thread ID in _before_request in addition to the span. Then in _teardown_request check that it's the same thread for that span before calling activation.__exit__

Fixes #1653
Fixes #1551

Type of change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)

How Has This Been Tested?

Please describe the tests that you ran to verify your changes. Provide instructions so we can reproduce. Please also list any relevant details for your test configuration

The bug is reproducible with this simple Flask app:

from concurrent.futures import ThreadPoolExecutor, as_completed
from random import randint

from flask import Flask, copy_current_request_context

from opentelemetry.instrumentation.flask import FlaskInstrumentor

app = Flask(__name__)

def do_random_stuff():
    @copy_current_request_context
    def inner():
        return randint(0, 100)
    return inner

@app.route("/")
def hello_world():
    count = 4
    executor = ThreadPoolExecutor(count)
    futures = []
    for _ in range(count):
        futures.append(executor.submit(do_random_stuff()))

    for future in as_completed(futures):
        print(future.result())

    return "Hello, World!"

FlaskInstrumentor.instrument_app(app)

When run with opentelemetry-instrument --traces_exporter console --metrics_exporter console flask run and accessed using HTTP it results in

    self.app.do_teardown_request(exc)
  File "~/otelbug/env/lib/python3.9/site-packages/flask/app.py", line 2373, in do_teardown_request
    self.ensure_sync(func)(exc)
  File "~/otelbug/env/lib/python3.9/site-packages/opentelemetry/instrumentation/flask/__init__.py", line 441, in _teardown_request
    activation.__exit__(None, None, None)
  File "/Library/Frameworks/Python.framework/Versions/3.9/lib/python3.9/contextlib.py", line 126, in __exit__
    next(self.gen)
ValueError: generator already executing

After the change, the issue no longer appears and the trace appears in the console.

Does This PR Require a Core Repo Change?

• No.

Checklist:

See contributing.md for styleguide, changelog guidelines, and more.

• Followed the style guidelines of this project
• Changelogs have been updated
• Unit tests have been added
• Documentation has been updated

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: hangonlyra (de1d1b7)

[shalevr]

Please add a changelog entry

[hangonlyra]

Please add a changelog entry

Updated. Thanks for the help.

[shalevr]

I am not familiar with flask requests,
Is there a legitimate way that the request starts with some thread and ends up in another thread?

[hangonlyra]

I am not familiar with flask requests, Is there a legitimate way that the request starts with some thread and ends up in another thread?

I can't think of any legitimate reason for that because it won't serve any useful purpose. It's definitely possible but it won't make any sense.

[matthewgrossman]

my approval in this repo doesn't mean anything, but just to echo from the initial issue that this will solve my issue

fixes #1551

[aabmass]

Thanks for the informative PR description! I'm not super familiar with this but the solution makes sense to me after reading through.

How difficult would it be to add a unit test for this?

[hangonlyra]

@aabmass Good idea. Added new test case TestMultiThreading.test_multithreaded

[aabmass]

Looks great! Thanks again

[srikanthccv]

Thanks