My cloud function's deployment fails undeterministically with an error

The same deployment, when I re-deploy it, has some non-deterministic chance to fail. Why would it happen and how do I ensure it's deployed correctly every time?

Hey there 👋 I have been trying to reproduce this for a long time now 😔 Do you have any insights into why this might be hapenning? How often, and possibly what pieces of code could cause it?

So far I know it's random, and only in Python

Not much insight, except that it came to my awareness only when I started using context.log inside my Python program.

Also, ChatGPT said it has something to do with a MySQL database given the error message - maybe something on Appwrite's backend?

Ah, another thing: in my deployment script, I only have pip install -r requirements.txt. The failure happens after the requirements are installed.

Correct, problem is with storing logs. Allowing this on database-level could introduce security issue with some null values. Database team asked me to try to reproduce and see what log exactly is problematic, and do some encoding instead.

What are you putting into context log?

This sounds more reasonable to me. Something internally is doing some native logs, which are then being sent as logs to DB. Do you have any insight on which lib it could be?

I have the following libs in the requirements file:

appwrite==4.0.0
openai==1.3.2
pinecone-client==2.2.4
python-dotenv==1.0.0
google-api-python-client==2.108.0
google-auth-httplib2==0.1.1
google-auth==2.23.4
tiktoken==0.5.1

One way to go might be to create a requirements file on your end with those libs and just spam deployments until it happens?

I don't think it's the context.log statement inside the code - I mean, my code shouldn't run on deployment time, right?

Oh, ooooh it's deployment. Usually people report this with executions

Yes, it's the deployment, not the execution

Ill try your requirements and see if I can reproduce locally

Let me know if you find anything!