Pydantic V2 slower on AWS lambda

[matthieudesprez]

Initial Checks

• I confirm that I'm using Pydantic V2

Description

Hello,

We've migrated from v1 to v2 in our codebase, and we've experienced slower execution times on AWS lambda, along with memory consumption increase.

I'm struggling to reproduce this locally, so for now I thought I could already share it here in case someone has a similar issue.

It doesn't seem to be related to cold start, as even warm lambdas are affected, but it looks like the most "wasted" time happens during the beginning of the execution. It takes 20s to do something I still can't figure.

Example Code

No response

Python, Pydantic & OS Version

pydantic version: 2.1.1
        pydantic-core version: 2.4.0
          pydantic-core build: profile=release pgo=false mimalloc=true
                 install path: .../lib/python3.8/site-packages/pydantic
               python version: 3.8.12 (default, Sep  5 2022, 15:57:41)  [Clang 13.1.6 (clang-1316.0.21.2.3)]
                     platform: ...
     optional deps. installed: ['email-validator', 'typing-extensions']

[adriangb]

Slow cold starts are a known problem.
Memory usage was a problem but should have been solved by pydantic/pydantic-core#900.

I see you're using pydantic-core 2.4.0. Can you try updating to the latest pydantic (v2.3.0 as or today) and hence the latest pydantic-core?

[matthieudesprez]

Thanks for your answer. It looks better indeed since upgrading to the last version.

Can we do something in order to optimise the initialisation time ?

[adriangb]

Just curious, are you using FastAPI?

[matthieudesprez]

Not on this lambda, it's actually triggered by Appsync as a resolver for a GraphQL Query, but it does use 5 levels depth nested models.

Last update seems to work better for now, so I'm gonna close this issue, thanks for your help.