template.yml

AWSTemplateFormatVersion: 2010-09-09
Description: >-
  page-capture-s3

Transform:
- AWS::Serverless-2016-10-31

Parameters:
  CaptureUrl:
    Type: String
    Description: URL of page to capture
    Default: https://www.bu.edu
  S3BucketName:
    Type: String
    Description: Name of the S3 bucket to deploy captured files
    Default: bu-static-page-bucket
  S3Path:
    Type: String
    Description: Path within the S3 bucket for deployed files
  SubdirPrefix:
    Type: String
    Description: Prefix of the S3 location (optional)
    Default: home

Resources:
  # SQS Queue to buffer page capture requests from the API Gateway.
  # Page captures can be long running jobs, and require a queue to receive and process API Gateway requests within their 30 second timeout.
  CaptureQueue:
    Type: AWS::SQS::Queue
    Properties:
      VisibilityTimeout: 230 # Set a high visibility timeout to give page captures at least a minute or so to complete
  
  PageCaptureFunction:
    Type: AWS::Serverless::Function
    Properties:
      Description: Lambda function to run the page capture and upload to S3. Runs when a request comes into the Capture Queue.
      Runtime: nodejs14.x
      Handler: src/page-capture.pageCapture
      Environment:
        Variables:
          CAPTURE_URL: !Ref CaptureUrl
          S3_BUCKET_NAME: !Ref S3BucketName
          S3_PATH: !Ref S3Path
          SUBDIR_PREFIX: !Ref SubdirPrefix
      Events:
        # This property associates this Lambda function with the SQS queue defined above, so that whenever the queue
        # receives a message, the Lambda function is invoked
        SQSQueueEvent:
          Type: SQS
          Properties:
            Queue: !GetAtt CaptureQueue.Arn
            BatchSize: 1
        # This event defines an hourly CloudWatch schedule
        CaptureHourlySchedule:
          Type: Schedule
          Properties:
            Schedule: 'rate(1 hour)'
            Description: Runs a capture every hour once enabled
            Enabled: false
      MemorySize: 512
      Timeout: 200 # must be within the SQS Visibility Timeout
      Policies:
        - AWSLambdaBasicExecutionRole
        - S3CrudPolicy:
            BucketName: !Ref S3BucketName

  # API Gateway endpoint secured with an API Key.  Anyone with the URL and API key can request new page captures.
  TriggerEndpoint:
    Type: AWS::Serverless::Api
    Properties:
      StageName: Prod
      Auth:
        ApiKeyRequired: true
        UsagePlan:
          CreateUsagePlan: PER_API
          UsagePlanName: GatewayAuthorization
          # Add an upper bound to monthly invocations just in case.
          Quota:
            Limit: 1000
            Period: MONTH

  # Simple Lambda function that relays requests from the API Gateway endpoint to SQS
  TriggerFunction:
    Type: AWS::Serverless::Function
    Properties:
      Description: Trigger to pass the API gateway event into SQS
      Runtime: nodejs14.x
      Handler: src/trigger.captureTriggerHandler
      Policies:
        SQSSendMessagePolicy: 
          QueueName: !GetAtt CaptureQueue.QueueName
      Environment:
        Variables:
          SQSqueueName: !Ref CaptureQueue
      Events:
        ApiKey:
          Type: Api
          Properties:
            RestApiId:
              Ref: TriggerEndpoint
            Path: /
            Method: get

Outputs:
  ApiGateway:
    Description: "The trigger URL is:"
    Value: !Sub "https://${TriggerEndpoint}.execute-api.${AWS::Region}.amazonaws.com/Prod/"
  ApiKey:
    Description: "You can find your API Key in the AWS console: (Put in the request HEADER as 'x-api-key')"
    Value: !Sub "https://console.aws.amazon.com/apigateway/home?region=${AWS::Region}#/api-keys/${TriggerEndpointApiKey}"