Enable run_task command to wait until the executions ends and return the exit code

[gonrial]

Describe Request:

Add the capability to wait until the end of an ECS task execution. Additionally, add the option to use the exit code as a way to track success or failure of the step.

The expected behaviors should be similar to the verify_revision_is_deployed parameter in update_service command. The difference with a service is that we know this task will end, and there is a different way to request the task status.

Default behavior (current one): no wait. Similar to fire-and-forget.

Examples:

description: Start the run of an ECS task on Fargate.
usage:
  version: 2.1
  orbs:
    aws-ecs: circleci/aws-ecs@6.0.0
    aws-cli: circleci/aws-cli@5.1.0
  jobs:
    run_task:
      docker:
        - image: cimg/python:3.10
      steps:
        - aws-cli/setup:
            profile_name: "OIDC-USER"
            role_arn: "arn:aws:iam::123456789012:role/VALID_OIDC_ECS_ROLE"
        - aws-ecs/run_task:
            cluster: cluster1
            region: us-east-1
            task_definition: myapp
            subnet_ids: '$SUBNET_ONE, $SUBNET_TWO'
            security_group_ids: $SECURITY_GROUP_IDS
            wait_for_task_stopped: true
  workflows:
    run_task:
      jobs:
        - run_task

Possible implementation using bash:

#!/bin/sh
AWS_PROFILE=default
ECS_CLUSTER=cluster_name
CONTAINER_NAME=migrate

# Run task and get its arn
# NOTE: many of the necessary cli inputs have been omitted here
TASK_ARN=$(aws ecs run-task \
  --cluster $ECS_CLUSTER \
  --query 'tasks[].taskArn' \
  --output text)

# Wait for ecs task to stop
aws ecs wait tasks-stopped \
  --cluster $ECS_CLUSTER \
  --tasks $TASK_ARN

# Get exit code
TASK_EXIT_CODE=$(aws ecs describe-tasks \
  --cluster $ECS_CLUSTER \
  --tasks $TASK_ARN \
  --query "tasks[0].containers[?name=='$CONTAINER_NAME'].exitCode" \
  --output text)

echo "The $TASK_ARN ran in ECS cluster $ECS_CLUSTER and its $CONTAINER_NAME returned exit code $TASK_EXIT_CODE"

# exit with the same code
exit $TASK_EXIT_CODE

Warning

From the AWS ecs wait tasks-stopped command documentation:

it will poll every 6 seconds until a successful state has been reached. This will exit with a return code of 255 after 100 failed checks.

The AWS CLI do not offer parameters to control the poll interval nor the number of poll attempts.

Supporting Documentation Links:

• https://docs.aws.amazon.com/cli/latest/reference/ecs/wait/tasks-stopped.html
• https://docs.aws.amazon.com/cli/latest/reference/ecs/describe-tasks.html

[marboledacci]

This feature is something to be on the run_task command or in the update_service?

[gonrial]

run_task command