beespinosa / Klayers

Python Packages as AWS Lambda Layers

Keith's Layers (Klayers)

🐍 A collection of Python Packages as AWS Lambda(λ) Layers 🐍

• List of ARNs
• Python Packages
• Using the Layers
  • Option 1: Using the Console
  • Option 2: Download copy of layer
  • Option 3: Using Serverless Framework
  • Option 4: Using AWS Serverless Application Model (SAM)
  • Option 5: Using Terraform with the Klayer provider
• Status of layers
• Layer expiry
• Architecture Diagram
• API
  • Get latest ARN for all packages in region
  • Get all ARNs for specific package in region
• Special Thanks
• Asking for additional layers

List of ARNs

List of the latest layer version arns are available by region:

• Python 3.8
• Python 3.9

Note: We have deprecated layers for python3.6 and python 3.7, please use these newer versions that are built for python 3.8 and python 3.9 going forward.

Python Packages

For the full list of Python packages, refer to the following, feel free to make a pull requests modifying the files below to requests for newer python versions.

• Python 3.8
• Python 3.9

Using the Layers

You can use the layers anyway you see fit, here are 4 options based on what method you use to deploy your lambda functions:

Option 1: Using the Console

Add the arn directly from the console, by selecting Layers->Add a Layer->Specify an Arn:

Option 2: Download copy of layer

Use the Get Layer Version by ARN in python or aws-cli command which will provide an S3 location to download the layer as a zip.

Note: You can only get layers from the specific region your client is configured for, otherwise you'll get a AccessDeniedException error.

Option 3: Using Serverless Framework

You can include layers in your deployments, by utilizing the layers property at the function level, and setting it to the arn of your choice. You must use layers from the same region as your function:

check:
  handler: 02_pipeline/check.main
  description: Checks for package on PyPi via the API
  runtime: python3.9
  timeout: 30
  memorySize: 256
  layers:
  - arn:aws:lambda:${self:provider.region}:113088814899:layer:Klayers-python37-packaging:1
  - arn:aws:lambda:${self:provider.region}:113088814899:layer:Klayers-python38-aws-lambda-powertools:23

Option 4: Using AWS Serverless Application Model (SAM)

Using AWS SAM, you can include layers in your serverless applications.You must use layers from the same region as your function:

ServerlessFunction:
  Type: AWS::Serverless::Function
  Properties:
    CodeUri: .
    Handler: my_handler
    Runtime: Python3.9
    Layers:
        - arn:aws:lambda:ap-southeast-1:113088814899:layer:Klayers-p39-packaging:1

Option 5: Using Terraform with the Klayer provider

Using Terraform, you can use the terraform-provider-klayer. The provider uses the API to enable your functions to always reference the latest layer versions.

terraform {
  required_providers {
    klayers = {
      version = "~> 1.0.0"
      source  = "ldcorentin/klayer"
    }
    aws = {
      source  = "hashicorp/aws"
      version = "~> 3.0"
    }
  }
}

provider "aws" {
  region = "us-east-1"
}

data "klayers_package_latest_version" "pandas" {
  name   = "pandas"
  region = "us-east-1"
}

resource "aws_lambda_function" "test_lambda" {
  filename      = "index.zip"
  function_name = "klayers-test"
  role          = aws_iam_role.iam_for_lambda.arn
  handler       = "index.handler"
  runtime       = "python3.9"
  layers = [
    data.klayers_package_latest_version.pandas.arn
  ]
}

Status of layers

Layers are built with the latest package version at 2am UTC on the first day of the Month.

If there's a critical issue with a package, which needs a new version of it published, raise an issue for the package, and I'll try my best to perform the deploy.

Layer expiry

Some layer versions will have a expiry_date field. This is the date for when the layers will be deleted.

In general, layers are scheduled for deletion 365 days after a new layer version has been published for that package. If you use that latest version of a layer, you're guaranteed at least 365 days before the layer is deleted.

All functions deployed with a layer will still work indefinitely, but you won't be able to deploy new functions referencing a deleted layer version.

Architecture Diagram

API

API v2 has been released. Please use v2 going forward, as we'll end v1 by June-2022.

All API calls are http-based, and work only with https (TLS1.2 and above). The API is heavily cached, so there could be minor delays in updates.

Get latest ARN for all packages in region

Returns data on the latest layer for a all packages in a specific {region} for your {python_version}. Defaults to json, but both csv and html can be returned as well.

https://api.klayers.cloud/api/v2/{python_version}/layers/latest/{region}/{format}

example:

• https://api.klayers.cloud/api/v2/p3.9/layers/latest/us-east-1/
• https://api.klayers.cloud/api/v2/p3.8/layers/latest/ap-southeast-1/
• https://api.klayers.cloud/api/v2/p3.9/layers/latest/ap-southeast-1/csv
• https://api.klayers.cloud/api/v2/p3.9/layers/latest/ap-southeast-1/html

Get all ARNs for specific package in region

Returns data on the all layers (latest and deprecated) for a specific {package} in a specific {region} for your {python_version}

https://api.klayers.cloud/api/v2/{python_version}/layers/{region}/{package}

example:

• https://api.klayers.cloud/api/v2/p3.8/layers/us-east-1/requests
• https://api.klayers.cloud/api/v2/p3.9/layers/ap-southeast-1/boto3

Special Thanks

• Chahna107
• ~ Dependencies scanned by PyUp.io ~
• @ldcorentin for making the awesome klayers-provider

Asking for additional layers

If you would like a new package to be made a layer, raise a pull request modifying the pipeline/config/packages_p39.csv or pipeline/config/packages_p38.csv file (depending on which version of python you use). By default, please request for the latest version of python unless you absolutely need an older version.