Skip to content

0002-lambda-deployment.md

Latest commit

 

History

History
87 lines (69 loc) · 3.71 KB

0002-lambda-deployment.md

File metadata and controls

87 lines (69 loc) · 3.71 KB
  • Start Date: 2020-12-21
  • Target Version: 1.0.0

Summary

This RFC proposes adding a new service deployment from PipeCD: AWS Lambda deployment. Similar to the current Google cloud run deployment.

Motivation

PipeCD aims to support wide range of deployable services, currently Terraform deployment and Cloud Run deployment are supported. Lambda deployment is the current missing piece for PipeCD's purpose.

Detailed design

Note:

  • Since the Lambda function as container image feature was available on AWS recently, we plan to implement PipeCD Lambda deployment based on that update.
  • To simplify the implementation, the initial release of this feature will only support container images from ECR ( the AWS container registry ).

Usage

The deployment configuration is used to customize the way to do the deployment. In case of Lambda function deployment, current common stage options (WAIT, WAIT_APPROVAL, ANALYSIS) are all inherited, besides with the stages for Lambda deployment itself: LAMBDA_CANARY_ROLLOUT, LAMBDA_TRAFFIC_ROUTING.

# https://docs.aws.amazon.com/lambda/latest/dg/configuration-versions.html
apiVersion: pipecd.dev/v1beta1
kind: LambdaApp
spec:
  input:
    image: xxxxxxxxxxxx.dkr.ecr.ap-northeast-1.amazonaws.com/demoapp:v0.0.1
  pipeline:
    stages:
      # Deploy workloads of the new version.
      # But this is still receiving no traffic.
      - name: LAMBDA_CANARY_ROLLOUT
      # Change the traffic routing state where
      # the new version will receive the specified percentage of traffic.
      # This is known as multi-phase canary strategy.
      - name: LAMBDA_TRAFFIC_ROUTING
        with:
          newVersion: 10
      # Optional: We can also add an ANALYSIS stage to verify the new version.
      # If this stage finds any not good metrics of the new version,
      # a rollback process to the previous version will be executed.
      - name: ANALYSIS
      # Change the traffic routing state where
      # thre new version will receive 100% of the traffic.
      - name: LAMBDA_TRAFFIC_ROUTING
        with:
          newVersion: 100

Architecture

Just as current Cloud Run but under pkg/cloudprovider/lambda package.

Alternatives

Lambda function could also be deployed via source code by 2 steps method:

  1. Configure piped to be able to clone the source of Lambda function ( on the same repo which deployment be handled by piped or via remote git repo ).
  2. Compress the source code and deploy with SAM or using aws-sdk ( piped handle the deployment ).

The deployment configuration sample as below:

In case of source code for Lambda function is on the same repo handled by Piped

apiVersion: pipecd.dev/v1beta1
kind: LambdaApp
spec:
  input:
    # Lambda code sourced from the same Git repository.
    path: lambdas/helloworld

In case of using difference git repo for lambda function source versioning

apiVersion: pipecd.dev/v1beta1
kind: LambdaApp
spec:
  input:
    # Lambda code sourced from another Git repository.
    git: git@github.com:org/source-repo.git
    path: lambdas/helloworld
    ref: v1.0.0

Unresolved questions

Currently, PipeCD team plans to implement this feature using the newly arrived Lambda function as container images feature of AWS ( as mentioned on #Detailed Design ). Since this feature is new ( even on AWS ), not all languages are supported to be deployed as container images currently.