Add timeout option

[clareliguori]

wait-for-services-stable currently only waits 10 minutes, which might not be long enough for all deployments

[amazingguni]

I need this too 😄

[clareliguori]

@amazingguni how long of a timeout do you need?

[amazingguni]

@clareliguori

In my case, it should be more than 15 minutes.
But I think it's better to support a parameter to set the timeout like circleci orbs aws/ecs.
(max-poll-attempts , 'poll-interval')
https://circleci.com/orbs/registry/orb/circleci/aws-ecs

echo "Verifying that << parameters.task-definition-arn >> is deployed.."
attempt=0
while [ "$attempt" -lt << parameters.max-poll-attempts >> ]
do
    DEPLOYMENTS=$(aws ecs describe-services \
        --cluster << parameters.cluster-name >> \
        --services ${SERVICE_NAME} \
        --output text \
        --query 'services[0].deployments[].[taskDefinition, status]')
    NUM_DEPLOYMENTS=$(aws ecs describe-services \
        --cluster << parameters.cluster-name >> \
        --services ${SERVICE_NAME} \
        --output text \
        --query 'length(services[0].deployments)')
    TARGET_REVISION=$(aws ecs describe-services \
        --cluster << parameters.cluster-name >> \
        --services ${SERVICE_NAME} \
        --output text \
        --query "services[0].deployments[?taskDefinition==\`<< parameters.task-definition-arn >>\` && runningCount == desiredCount && (status == \`PRIMARY\` || status == \`ACTIVE\`)][taskDefinition]")
    echo "Current deployments: $DEPLOYMENTS"
    if [ "$NUM_DEPLOYMENTS" = "1" ] && [ "$TARGET_REVISION" = "<< parameters.task-definition-arn >>" ]; then
        echo "The task definition revision $TARGET_REVISION is the only deployment for the service and has attained the desired running task count."
        exit 0
    else
        echo "Waiting for revision << parameters.task-definition-arn >> to reach desired running count / older revisions to be stopped.."
        sleep << parameters.poll-interval >>
    fi
    attempt=$((attempt + 1))
done
echo "Stopped waiting for deployment to be stable - please check the status of << parameters.task-definition-arn >> on the AWS ECS console."

[clareliguori]

@amazingguni Yep makes sense to have a configurable timeout, and probably also to raise the default timeout :)