Python + Poetry in serverless - A few things to make sure you have in your serverless.yml file

.gitignore

# Distribution / packaging
.Python
.DS_Store
env/
build/
develop-eggs/
dist/
downloads/
eggs/
.eggs/
lib/
lib64/
parts/
sdist/
var/
*.egg-info/
.installed.cfg
*.egg
.venv # venv from poetry w/ 

# Serverless directories and stuff 
.serverless
_dev
requirements.txt

poetry.toml

[virtualenvs]
in-project = true

serverless.yml

frameworkVersion: '2'

custom:
  pythonRequirements:
    dockerizePip: non-linux
  scripts:
    hooks:
      'package:initialize': poetry export -f requirements.txt > requirements.txt

provider:
  name: aws
  runtime: python3.8
  lambdaHashingVersion: 20201221
  tracing:
    lambda: true
package:
 # include:
 #   - include-me.py
 #   - include-me-dir/**
 exclude:
   - node_modules
   - .env
   - package-lock.json
   - package.json
   - poetry.toml
   - poetry.lock
   - .flake8
   - .git
   - .gitignore
   - .pylintrc
   - _data
   - docker-compose.yml
   - .github
   - .pytest_cache
   - .venv
   - tests
useDotenv: true
plugins:
  - serverless-python-requirements

So I use Poetry because it's awesome, but the way serverless is setup, it still looks for a good 'ol requirements.txt... so... easiest way around this is twofold:

• Add a custom.scripts.hooks.package:initialize script to dump out your requirements.txt from your poetry setup
• Add requirements.txt to your .gitignore file

I've also thrown in some excludes to make sure my local stuff doesn't overload the lambda env (since there's a 75MB limit)...

Generally... I tend to:

• Set up poetry with the in-project set to true which generates a .venv directory in the project (I like to keep things together)
• given that, my .gitignore file has some extra stuff in it to prevent uploading common stuff that shouldn't go to the repo (dependencies and such...)