DOP-C02 Practice Question 32: ECS Deployment Circuit Breaker

The deployment circuit breaker is described as "the rolling update mechanism that determines if the tasks reach a steady state," and it "has an option that will automatically roll back a failed deployment to the deployment that is in the COMPLETED state." When the failure count reaches the threshold, the deployment moves to FAILED, stops launching tasks, and, with rollback enabled, ECS starts a new deployment from the last completed one. It is supported only for services using the rolling update deployment controller, which is what this service already uses, so nothing about the deployment model or the load balancer has to change.

Option A protects capacity but never ends the bad deployment. With no healthy new tasks, ECS simply keeps trying within those percentages for as long as the deployment lasts, which is the behavior being complained about.

Option B makes ECS notice an unhealthy task faster, and therefore churn faster. Health check tuning has no concept of a failed *deployment*, so the deployment still never fails and never rolls back.

Option D would give automatic rollback, but it changes the deployment controller and an ECS blue/green deployment requires two target groups, both explicitly ruled out by the stem.

Reference: How the Amazon ECS deployment circuit breaker detects failures