ecs fargate Deployment failure detection

• outline

When the fail count exceeds a certain number of times, it returns to the previous deployment state. When the ECS Service is first created, there is no previous deployment state, so it is deleted.

• Deployment failure detection

1. When initial configuration of ECS is done through Web Console, the settings below are default(Rollback on failures). Due to the nature of the initial creation, there is a high probability that a deployment will fail, and then the service is deleted if it exceeds a certain number of times. Therefore, I recommend testing without the settings below.

1. This setting is useful not only because it is a circuit breaker, but also because it prevents infinite restart in the event of an unexpected deployment failure.

• configuration

Deployment failure detection
\> Use the Amazon ECS deployment circuit breaker
\> Rollback on failures

• note

When a service is automatically deleted and re-created, it remains in the cloudformation stack with the same name as before, so the stack must be deleted to create a service with the same name, again.