- 最新
- 投票最多
- 评论最多
If you are not using Circuit Breaker or CloudWatch Alarms to roll back your Deployment on failed ECS Tasks during the deployment you have to deploy a new Deployment with new Task Definition which will pass health checks, or revert the Task Definition version to the last known working version.
According to the https://aws.amazon.com/blogs/containers/automate-rollbacks-for-amazon-ecs-rolling-deployments-with-cloudwatch-alarms/ blog post: "By default, when the tasks in a service are unable to reach a healthy state, the scheduler applies the service throttling logic to retry the deployments and it required manual intervention to cancel the deployments after a certain number of failed attempts."
ECS Deployment "stuck" in the "In progress" state is expected behavior.
相关内容
- AWS 官方已更新 1 年前
- AWS 官方已更新 2 年前