bvsam / aws-lambda-deploy

Fast model deployment on AWS Lambda

Bentoctl AWS Lambda deployment

Bentoctl is a CLI tool for deploying your machine-learning models to any cloud platforms and serving predictions via REST APIs. It built on top of BentoML: the unified model serving framework, and makes it easy to bring any BentoML packaged model to production.

This repo contains the Bentoctl AWS Lambda deployment operator. This operator defines the terraform configuration for deploying the Lambda function and how to build docker image that's compatible with AWS Lambda.

Note: This operator is compatible with BentoML version 1.0.0 and above. For older versions, please switch to the branch pre-v1.0 and follow the instructions in the README.md.

Table of Contents

• Prerequisites
• Quickstart with bentoctl
• Configuration options

Prerequisites

1. Bentoml - BentoML version 1.0 and greater. Please follow the Installation guide.
2. Terraform - Terraform is a tool for building, configuring, and managing infrastructure.
3. AWS CLI installed and configured with an AWS account with permission to the Cloudformation, Lamba, API Gateway and ECR. Please follow the Installation guide.
4. Docker - Install instruction: https://docs.docker.com/installll instruction: https://www.terraform.io/downloads.html
5. A built Bento project. For this guide, we will use the Iris classifier bento from the BentoML quickstart guide.

Quickstart with bentoctl

Bentoctl is a CLI tool that you can use to deploy bentos to Lambda. It helps in configuring and managing your deployments super easy.

1. Install bentoctl via pip

   $ pip install bentoctl
   

2. Install AWS Lambda operator

   Bentoctl will install the official AWS Lambda operator and its dependencies.

   $ bentoctl operator install aws-lambda
   

3. Initialize deployment with bentoctl

   Follow the interactive guide to initialize deployment project.

   $ bentoctl init
   
   Bentoctl Interactive Deployment Config Builder
   
   Welcome! You are now in interactive mode.
   
   This mode will help you setup the deployment_config.yaml file required for
   deployment. Fill out the appropriate values for the fields.
   
   (deployment config will be saved to: ./deployment_config.yaml)
   
   api_version: v1
   name: demo
   operator: aws-lambda
   template: terraform
   spec:
       region: us-west-1
       timeout: 10
       memory_size: 512
   filename for deployment_config [deployment_config.yaml]:
   deployment config file exists! Should I override? [Y/n]:
   deployment config generated to: deployment_config.yaml
   ✨ generated template files.
     - bentoctl.tfvars
     - main.tf

4. Build and push AWS Lambda comptable docker image to registry

   Bentoctl will build and push the Lambda compatible docker image to the AWS ECR repository.

   bentoctl build -b iris_classifier:latest -f deployment_config.yaml
   
   Step 1/20 : FROM bentoml/bento-server:1.0.0a7-python3.7-debian-runtime
   ---> dde7b88477b1
   Step 2/20 : ARG UID=1034
   ---> Running in b8f4ae1d8b08
   ---> e6c313c8d9ea
   Step 3/20 : ARG GID=1034
   ....
   Step 20/20 : ENTRYPOINT [ "/opt/conda/bin/python", "-m", "awslambdaric" ]
   ---> Running in 4e56057f3b18
   ---> dca82bca9034
   Successfully built dca82bca9034
   Successfully tagged aws-lambda-iris_classifier:btzv5wfv665trhcu
   🔨 Image build!
   The push refers to repository [192023623294.dkr.ecr.us-west-1.amazonaws.com/quickstart]
   btzv5wfv665trhcu: digest: sha256:ffcd120f7629122cf5cd95664e4fd28e9a50e799be7bb23f0b5b03f14ca5c672 size: 3253
   32096534b881: Pushed
   f709d8f0f57d: Pushed
   7d30486f5c78: Pushed
   ...
   c1065d45b872: Pushed
   🚀 Image pushed!
   ✨ generated template files.
     - bentoctl.tfvars
   The push refers to repository [192023623294.dkr.ecr.us-west-1.amazonaws.com/quickstart]

5. Apply Deployment with Terraform

   1. Initialize terraform project

      terraform init

   2. Apply terraform project to create Lambda deployment

      terraform apply -var-file=bentoctl.tfvars -auto-approve
      
      aws_iam_role.lambda_exec: Creating...
      aws_apigatewayv2_api.lambda: Creating...
      aws_apigatewayv2_api.lambda: Creation complete after 1s [id=ka8h2p2yfh]
      aws_cloudwatch_log_group.api_gw: Creating...
      aws_cloudwatch_log_group.api_gw: Creation complete after 0s [id=/aws/api_gw/quickstart-gw]
      aws_apigatewayv2_stage.lambda: Creating...
      aws_iam_role.lambda_exec: Creation complete after 3s [id=quickstart-iam]
      aws_iam_role_policy_attachment.lambda_policy: Creating...
      aws_lambda_function.fn: Creating...
      aws_apigatewayv2_stage.lambda: Creation complete after 2s [id=$default]
      aws_iam_role_policy_attachment.lambda_policy: Creation complete after 1s [id=quickstart-iam-20220414203448384500000001]
      aws_lambda_function.fn: Still creating... [10s elapsed]
      aws_lambda_function.fn: Still creating... [20s elapsed]
      aws_lambda_function.fn: Still creating... [30s elapsed]
      aws_lambda_function.fn: Still creating... [40s elapsed]
      aws_lambda_function.fn: Creation complete after 41s [id=quickstart-function]
      aws_lambda_permission.api_gw: Creating...
      aws_cloudwatch_log_group.lg: Creating...
      aws_apigatewayv2_integration.lambda: Creating...
      aws_lambda_permission.api_gw: Creation complete after 0s [id=AllowExecutionFromAPIGateway]
      aws_cloudwatch_log_group.lg: Creation complete after 0s [id=/aws/lambda/quickstart-function]
      aws_apigatewayv2_integration.lambda: Creation complete after 1s [id=8gumjws]
      aws_apigatewayv2_route.root: Creating...
      aws_apigatewayv2_route.services: Creating...
      aws_apigatewayv2_route.root: Creation complete after 0s [id=jjp5f23]
      aws_apigatewayv2_route.services: Creation complete after 0s [id=8n57a1d]
      
      Apply complete! Resources: 11 added, 0 changed, 0 destroyed.
      
      Outputs:
      
      base_url = "https://ka8h2p2yfh.execute-api.us-west-1.amazonaws.com/"
      function_name = "quickstart-function"
      image_tag = "192023623294.dkr.ecr.us-west-1.amazonaws.com/quickstart:btzv5wfv665trhcu"

6. Test deployed endpoint

   The iris_classifier uses the /classify endpoint for receiving requests so the full URL for the classifier will be in the form {EndpointUrl}/classify

   URL=$(terraform output -json | jq -r .base_url.value)classify
   curl -i \
     --header "Content-Type: application/json" \
     --request POST \
     --data '[5.1, 3.5, 1.4, 0.2]' \
     $URL
   
   HTTP/2 200
   date: Thu, 14 Apr 2022 23:02:45 GMT
   content-type: application/json
   content-length: 1
   apigw-requestid: Ql8zbicdSK4EM5g=
   
   0%

7. Delete deployment Use the bentoctl destroy command to remove the registry and the deployment

   bentoctl destroy -f deployment_config.yaml

Configuration options

• region: AWS region for Lambda deployment
• timeout: Timeout per request
• memory_size: The memory for your function, set a value between 128 MB and 10,240 MB in 1-MB increments