AlertResponder : Automatic Security Alert Response Framework By AWS Serverless Application Model

AlertResponder is an automatic security alert response framework by AWS Server less Application Model. It is a server less framework for automatic response of security alert.


AlertResponder receives an alert that is event of interest from security view point and responses the alert automatically. AlertResponder has 3 parts of automatic response.

  • Inspector investigates entities that are appeared in the alert including IP address, Domain name and store a result: reputation, history of malicious activities, associated cloud instance and etc.

    Following components are already provided to integrate with your AlertResponder environment. Also you can create own inspector to check logs that is stored into original log storage or log search system.
    • VirusTotalInspector
  • Reviewer receives the alert with result(s) of Inspector and evaluate severity of the alert.

    Reviewer should be written by each security operator/administrator of your organization because security policies are differ from organization to organization.
  • Emitter finally receives the alert with result of Reviewer’s severity evaluation. After that, Emitter sends external integrated system. E.g. PagerDuty, Slack, Github Enterprise, etc. Also automatic quarantine can be configured by AWS Lambda function.
    • GheReporter

Also Read – Grouper2 : To Find Vulnerabilities In AD Group Policy


  • Pull based correlation analysis
  • Alert aggregation
  • Pluggable Inspectors and Emitters

Getting Started

Please replace following variables according to your environment:

  • $REGION: Replace it with your AWS region. (e.g. ap-northeast-1)
  • $STACK_NAME: Replace it with CloudFormation stack name

$ curl -o alert_responder.yml https://s3-$$REGION/AlertResponder/templates/latest.yml

$ aws cloudformation deploy –template-file alert_responder.yml –stack-name $STACK_NAME –capabilities CAPABILITY_IAM


Architecture Overview:


  • awscli >= 1.16.20
  • Go >= 1.11
  • GNU automake >= 1.16.1

Deploy and Test

Deploy own AlertResponder stack

Prepare a parameter file, e.g. config.json and run make command.

$ cat config.json
“StackName”: “your-alert-responder-name”,
“TestStackName”: “your-test-stack-name”,
“CodeS3Bucket”: “your-some-bucket”,
“CodeS3Prefix”: “for-example-functions”,

“InspectionDelay”: “1”,
“ReviewDelay”: “10”
$ env AR_CONFIG=config.json make deploy

Deploy a test stack

After deploying AlertResponder, move to under tester directory and deploy a stack for testing.

$ cd tester/
$ make AR_CONFIG=../config.json deploy

You can see param.json that is created by script under tester directory after deploying.

$ cat params.json
“AccountId”: “214219211678”,
“Region”: “ap-northeast-1”,
“Inspector”: “slam-alert-responder-test-functions-Inspector-1OBGU89CT1P4B”,
“Reporter”: “slam-alert-responder-test-functions-Reporter-1NDHU0VDI8OPA”

Then, back to top level directory of the git repository and you can run integration test.

$ go test -v
=== RUN TestInvokeBySns
— PASS: TestInvokeBySns (3.39s)
ok 20.110s