Is there a way to roll back if there is a health check error after the service is normally deployed to ecs?

I deployed the web server through codedeploy.

The distribution process showed the result that it was normal.

However, after deployment, an error occurred in the health check of the load balancer.

The cause is that the web server is up, but the db connection is not connected, so the page does not appear, causing a health check error.

In this case, my ecs instance went down and up.

When I looked it up, it seemed that I could use a function called Deployment failure detection, but the function of the Amazon ECS deployment circuit breaker seemed useless if the service went up normally.

However, using cloudwatch alarms seemed appropriate for the situation I encountered.

I want to know if what I'm thinking is correct.