Static typing: The add_handler API from logger was removed

[alexandreczg]

Expected Behaviour

In version 2.4.0 adding a handler to the log was easy. I would expect adding a handler to my log to be as simple as addHandler.

Current Behaviour

Method doesn't exist

Code snippet

logger = logging.Logger()
logger.addHandler(additional_handler)

But that API was removed on version 2.15.0. I depend on it to be able to log to cloudwatch and other places. I was reading the changelog but could not find any reference to this.

the only way around of this is to mess with the internals:

logger._logger.addHandler(additional_handler)

### Possible Solution

Semver would dictate that the public API doesn't change with a minor version upgrade. Did I miss something something regarding this breaking change? Is there another workaround?

### Steps to Reproduce

Above

### Powertools for AWS Lambda (Python) version

latest

### AWS Lambda function runtime

3.10

### Packaging format used

PyPi

### 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

[alexandreczg]

Apologies, closing, cannot reproduce it again.

[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.

[heitorlessa]

Apologies, closing, cannot reproduce it again.

Hey @alexandreczg - did you get a runtime error or was a static checker error saying the method doesn't exist?

The former would be a regression tho I can't reproduce it either. If it's the latter, we can address that with a PR to implement it and proxy it to _logger.addHandler so mypy is happy.

This is the change I made that would cause the latter but not the former (tested with a stream handler + file log handler now).

Please let us know ;)

PS: No need to apologise, better safe than sorry and we appreciate your time flagging it very much!!

#2334

[alexandreczg]

Hey @heitorlessa, thank your for addressing this, commendable your dedication!

It was a typing error actually, not runtime. I use pyright/pylance (vscode). I freaked out since it had been a long hours day and when I was about to push my code I start getting that error. Couldn't really understand it since it had been working. I regressed to version 2.4.0 (it was my last stable version) but quickly realized that I had implemented the new batch partial processor (SQS). So, I went back to 2.16.0 and magically the linting error went way... I'm not sure what or how it all happened, but when I introspected the logger class, I couldn't find the addHandler method.

I'll keep your and this thread updated if something is to change now that I have pushed my code to AWS and using latest layer. 🤞

[alexandreczg]

@heitorlessa , okay in version 2.16.1 the issue is back

[heitorlessa]

@heitorlessa , okay in version 2.16.1 the issue is back

That's the piece I mentioned it's expected -- does it fail if you run it?

Pyright (what Pylance uses) can't infer through the object proxy so it believes it doesn't exist. At runtime, it works.

If you can confirm it works, feel free to make a PR adding a "addHandler" method that call self._logger.addHandler so Pylance will be happy in the next release - more than happy to accommodate that ;)