Run AWS step functions offline with Serverless!

This is a plugin for the Serverless Framework. It uses stepfunctions-localhost to emulate step functions with AWS' provided tool for local development.

• serverless >= v3.0.0
• The serverless-offline plugin
• The serverless-offline-lambda plugin
• The serverless-step-functions plugin
• Java Runtime Engine (JRE) version 6.x or newer

npm install serverless-step-functions-local -D

You'll need to add this plugin to your serverless.yml. The plugins section should look something like this when you're done:

plugins:
  ...
  - serverless-step-functions
  - serverless-step-functions-local
  - serverless-offline-lambda
  - serverless-offline
  ...

Then, add a new section to config with accountId and region parameters:

custom:
  stepFunctionsLocal:
    accountId: 101010101010
    region: us-east-1

Although not neccessary, it's strongly recomended to add the folder with the downloaded step function executables to .gitignore. By default, this path is ./.step-functions-local.

The plugin binds to port 8083, this cannot be changed. The plugin binds to port 8083, since stepfunctions-localhost doesn't support dynamic port, it can't be changed.

It also adds an environment variable for each created state machine that contains the ARN for it. These variables are prefixed by OFFLINE_STEP_FUNCTIONS_ARN_, so the ARN of a state machine named 'WaitMachine', for example could be fetched by reading OFFLINE_STEP_FUNCTIONS_ARN_WaitMachine.

(These go under custom.stepFunctionsLocal.)

• accountId (required) your AWS account ID
• region (required) your AWS region
• lambdaEndpoint (defaults to http://localhost:3000 from serverless offline) the endpoint for the lambda service. This plugin tries to use the configured value from serverless-offline
• path (defaults to ./.step-functions-local) the path to store the downloaded step function executables
• TaskResourceMapping allows for Resource ARNs to be configured differently for local development. (Serverless 3.x prefixes the lambda function names with the stack name and the stage)
• eventBridgeEvents allows sending EventBridge events on execution status changes
  • enabled (bool) enabled or disable this feature. Disabled by default.
  • endpoint Endpoint for sending events to eg. for serverless-offline-aws-eventbridge would be http://localhost:4010
• externalInstance allows to set a flag to use external instance of local step functions. Disabled by default.
• stepFunctionsEndpoint (defaults to http://localhost:8083) the endpoint for the step functions service

service: local-step-function

plugins:
  - serverless-step-functions
  - serverless-step-functions-local
  - serverless-offline-lambda
  - serverless-offline

provider:
  name: aws
  runtime: nodejs14.x # supported nodejs versions: https://docs.aws.amazon.com/lambda/latest/dg/lambda-runtimes.html


custom:
  stepFunctionsLocal:
    accountId: 101010101010
    region: us-east-1
    TaskResourceMapping:
      FirstState: arn:aws:lambda:us-east-1:101010101010:function:${self:service}-${opt:stage, self:provider.stage}-hello
      FinalState: arn:aws:lambda:us-east-1:101010101010:function:${self:service}-${opt:stage, self:provider.stage}-hello
    eventBridgeEvents:
      enabled: true
      endpoint: http://localhost:4010
  sqsUrl: http://localhost:4566/101010101010/example-queue

functions:
  hello:
    handler: handler.hello

stepFunctions:
  stateMachines:
    WaitMachine:
      definition:
        Comment: "An example of the Amazon States Language using wait states"
        StartAt: FirstState
        States:
          FirstState:
            Type: Task
            Resource: Fn::GetAtt: [hello, Arn]
            Next: send_message
          send_message:
            Type: Task
            Resource: arn:aws:states:::sqs:sendMessage
            Parameters:
              QueueUrl: ${self:custom.sqsUrl}
              "MessageBody.$": "$"
            Next: wait_using_seconds
          wait_using_seconds:
            Type: Wait
            Seconds: 10
            Next: FinalState
          FinalState:
            Type: Task
            Resource: Fn::GetAtt: [hello, Arn]
            End: true

To run your local step function, use this command:

aws stepfunctions --endpoint-url http://localhost:8083 start-execution --state-machine-arn arn:aws:states:us-east-1:101010101010:stateMachine:servicename-stagename-hello --input '{"hello": ["world"]}'

To run external step functions local using docker, use this command:

docker run \
    -p 8083:8083 \
    -e "AWS_ACCOUNT_ID=101010101010" \
    -e "AWS_DEFAULT_REGION=us-east-1" \
    -e "LAMBDA_ENDPOINT=http://localhost:4000" \
    amazon/aws-stepfunctions-local