After updating an ECS task definition with a new container image, an ECS service connected to an Application Load Balancer repeatedly stops and starts tasks because ALB health checks fail. What should a DevOps engineer check first to troubleshoot the failing deployment?

When an ECS service repeatedly stops and starts tasks due to ALB health check failures after a task definition update, it often indicates that the new container image or application within the container needs more time to initialize and become healthy. Increasing the ALB health check grace period allows newly launched tasks a longer duration to pass health checks before being marked unhealthy and stopped. This prevents premature termination of tasks that are still starting up. While ensuring the security group allows traffic is fundamental, if the service was previously working, a task definition update wouldn't inherently change security group rules. Increasing the minimum healthy percent would make the deployment more restrictive, potentially exacerbating the issue if tasks are failing health checks. Decreasing the health check interval would cause tasks to be marked unhealthy even faster, worsening the problem.