I added an ECS Fargate healthcheck to a new version of a task definition. The health check failed so the new tasks were never healthy, however, my old properly running task definitions were spun down before the new ones got healthy, which would have been downtime if we were live. How can I avoid this scenario? I want my new task to pass health checks before the old tasks are killed. This happened with

pulumi up

If this is part of an ECS service, set the following parameters to the ECS service resources:

deployment_maximum_percent=200
deployment_minimum_healthy_percent=100

This way the service will start a new task and have it healthy before it kills the previous one.

Thank you