Bug: AppConfig Provider Returning Wrong Configuration

[mekaeelkarim]

Expected Behaviour

When using the AppConfig Provider to fetch user specific configurations at runtime based on the identity of the requester that invokes our lambda function, we expect the correct configuration to be returned. Our AppConfig Application and Environment names are used to instantiate a global AppConfig Provider object at runtime, and the this provider is used to fetch the configuration specific to the requester multiple times throughout the execution of the request in that lambda container. We do not cache the profile and instead rely on the caching that is built into the AppConfig Provider.

Current Behaviour

What we are seeing is that when multiple requests are being made at the same time from different users (across different lambda invocations), the AppConfig provider will be called to fetch the configuration with the right username, but return the wrong configuration profile (it will be called with user A, but return user B's profile). We have tried turning off caching of the profiles and this did not help. All we are doing is calling that initialized AppConfigProvider with the config_name that is specific to each user when we need to fetch that profile. It is important to note that for each request, we only need to fetch one profile that is scoped to the user who made the request.

Code snippet

# Global AppConfig Provider Object
appconfig_client = AppConfigClient(
    application=APPLICATION_NAME,
    environment=ENVIRONMENT_NAME,
)

# Method used to fetch the Configuration
def get_config_as_dict(self, config_name: str) -> Dict[str, Any]:
     logger.info(f"Fetching configuration: {config_name}...")
     return cast(
         dict,
         self._app_config_client.get(
             name=config_name, transform="json", max_age=600
         ),
     )

Possible Solution

No response

Steps to Reproduce

Multiple requests attempting to fetch different configurations at the same time should be able to reproduce this error. The TPS for our service should not be more than 2-3 per second at most. Lambda containers being reused across requests most likely does play a part here.

Powertools for AWS Lambda (Python) version

2.17

AWS Lambda function runtime

3.8

Packaging format used

Lambda Layers

Debugging logs

No response

[boring-cyborg]

Thanks for opening your first issue here! We'll come back to you as soon as we can.
In the meantime, check out the #python channel on our Powertools for AWS Lambda Discord: Invite link

[leandrodamascena]

Hello @mekaeelkarim! Thank you for opening this issue. I was able to reproduce the error and will work to fix it. We will likely include this fix in our next release, which is scheduled for tomorrow.

[mekaeelkarim]

Hi @leandrodamascena, sounds good. Thanks for looking into this!

[github-actions]

⚠️COMMENT VISIBILITY WARNING⚠️

This issue is now closed. Please be mindful that future comments are hard for our team to see.

If you need more assistance, please either tag a team member or open a new issue that references this one.

If you wish to keep having a conversation with other community members under this issue feel free to do so.

[github-actions]

This is now released under 2.26.0 version!