Skip to content

AWS Lambda layer

Albert publishes a public AWS Lambda layer for each SDK release. Attach it to your function and import albert works without bundling the SDK or its dependencies in your deployment package.

Choosing a layer

One layer exists per Python runtime and architecture. Pick the one that matches your function's configuration:

Layer name Runtime Architecture
albert-python-py311-x86_64 python3.11 x86_64
albert-python-py311-arm64 python3.11 arm64
albert-python-py312-x86_64 python3.12 x86_64
albert-python-py312-arm64 python3.12 arm64
albert-python-py313-x86_64 python3.13 x86_64
albert-python-py313-arm64 python3.13 arm64
albert-python-py314-x86_64 python3.14 x86_64
albert-python-py314-arm64 python3.14 arm64

Regions

Lambda layers are region scoped: your function must be in the same region as the layer, or AWS reports a permissions error even though the layer is public. Every release is published to these regions by default:

  • us-west-2 (US West, Oregon)
  • us-east-1 (US East, N. Virginia)
  • eu-central-1 (Europe, Frankfurt)
  • eu-west-1 (Europe, Ireland)

Need the layer in another region? Open an issue with the region code and we can add it to this list, so it is published there for every release.

Available layer versions

AWS does not let other accounts browse or search our layers, so every public layer version is listed here. Filter by SDK version, Python runtime, architecture, and your function's region, then copy the ARN:

Loading layer versions. If this does not load, download lambda-layers.json directly.

The same list is published as JSON at https://docs.developer.albertinvent.com/albert-python/lambda-layers.json for scripts and infrastructure code. Entries are sorted newest SDK version first, so the first match is the latest layer:

curl -s https://docs.developer.albertinvent.com/albert-python/lambda-layers.json \
  | jq -r '[.layers[] | select(.python == "3.12" and .architecture == "x86_64" and .region == "us-west-2")][0].arn'

Each entry includes sdk_version, python, architecture, region, layer_name, layer_version, arn, created, and description.

Versioning

Each SDK release adds a new version to every layer. The layer version number is assigned by AWS and is not the SDK version. The table above maps one to the other, and each layer's description also records the SDK version it contains:

aws lambda get-layer-version-by-arn \
  --arn arn:aws:lambda:us-west-2:<albert-account-id>:layer:albert-python-py312-x86_64:3 \
  --query Description --output text

albert-python 1.19.0 | python3.12 | x86_64 | 2026-09-17T14:02:00Z | sha=abc1234

Every GitHub release also ends with a Lambda layers table listing the layer version ARNs published for that release. Pin your function to a specific layer version ARN and upgrade deliberately, the same way you would pin a package version.

Attaching the layer

aws lambda update-function-configuration \
  --function-name my-function \
  --layers arn:aws:lambda:us-west-2:<albert-account-id>:layer:albert-python-py312-x86_64:3

MyFunction:
  Type: AWS::Serverless::Function
  Properties:
    Runtime: python3.12
    Architectures: [x86_64]
    Layers:
      - arn:aws:lambda:us-west-2:<albert-account-id>:layer:albert-python-py312-x86_64:3

resource "aws_lambda_function" "my_function" {
  runtime       = "python3.12"
  architectures = ["x86_64"]
  layers        = ["arn:aws:lambda:us-west-2:<albert-account-id>:layer:albert-python-py312-x86_64:3"]
  # ...
}

What the layer contains

The layer installs albert and its dependencies (pandas, numpy, pydantic, requests, pyjwt, tenacity) under python/, which Lambda adds to sys.path. Do not bundle pandas or numpy in your own deployment package as well: duplicate copies waste space and can shadow the versions the SDK was built against.

Credentials work the same way as anywhere else. See Authentication for using a client credential token from an environment variable, which is the usual approach in Lambda.