AWS Messaging Plugin (SNS and SQS)

plugins:
  - name: asyncapi
  - name: aws-messaging
  - name: json

Use aws-messaging in place of the kafka plugin. One plugin covers both SNS and SQS, and Drift selects the right service per operation.

SNS and SQS support asynchronous interactions only - fire-and-forget flows where a message travels one way.

Drift determines the service per operation from the channel address, falling back to the server protocol.

Drift can both publish and consume on either service, so all three asynchronous execution modes are available whichever one your AsyncAPI document declares.

Use an ARN or an SQS queue URL as the AsyncAPI channel address:

channels:
  inventoryAdjusted:
    address: arn:aws:sqs:us-east-1:123456789012:inventory-adjusted

  orderNotifications:
    address: arn:aws:sns:us-east-1:123456789012:order-notifications

servers:
  aws:
    host: sqs.us-east-1.amazonaws.com
    protocol: sqs

Drift reads AWS credentials from standard environment variables and signs every request with SigV4.

export AWS_ACCESS_KEY_ID="AKIA..."
export AWS_SECRET_ACCESS_KEY="..."
export AWS_SESSION_TOKEN="..."      # only for temporary credentials

drift verify -f drift/inventory.testcases.yaml

Drift fails with a clear message if AWS_ACCESS_KEY_ID or AWS_SECRET_ACCESS_KEY is not set.

Do not commit credentials to your test case files. Use environment variables in local development and your CI system's secret store in pipelines. For general Drift authentication options, see Configuring Authentication.

Drift resolves the region in this order. The first match wins.

parameters:
  region: eu-west-2

Set an aws.endpoint-url parameter to send requests somewhere other than the public AWS endpoint. This is how you run against LocalStack or another emulator.

parameters:
  aws:
    endpoint-url: http://localhost:4566
  region: us-east-1

When no override is given, Drift builds the endpoint from the service and region as https://<service>.<region>.amazonaws.com.

services:
  localstack:
    image: localstack/localstack:latest
    ports:
      - "4566:4566"
    environment:
      SERVICES: sqs,sns

export AWS_ACCESS_KEY_ID=test
export AWS_SECRET_ACCESS_KEY=test
export AWS_REGION=us-east-1

aws --endpoint-url=http://localhost:4566 sqs create-queue \
  --queue-name inventory-adjusted

parameters:
  aws:
    endpoint-url: http://localhost:4566

Drift carries the correlation identifier as a message attribute on publish, and filters captured messages by that attribute.

Declare the location in your AsyncAPI message. This is required - it is what enables correlation.

components:
  messages:
    inventoryAdjusted:
      correlationId:
        location: "$message.header#/correlation-id"

The value is optional. Drift takes it from the AsyncAPI document, then from the correlation-id parameter, and generates a UUID when neither supplies one.

parameters:
  correlation-id: receive-inventory-adjusted-001

Set the value yourself when a probe command or an expectation needs the same identifier. Where you do set it, give each operation a distinct value.

Drift captures from SQS using long polling, which keeps latency low without busy-waiting.

parameters:
  timeout-ms: 10000

Allow more time than you would for a local Kafka broker. Network round trips to AWS add latency, and SQS visibility behavior can delay delivery.

Drift does not create queues or topics. Create every queue, topic, and probe queue in your AsyncAPI document before the run, using the AWS CLI, your infrastructure-as-code tool, or LocalStack setup scripts.

See Debugging Test Cases.