1 回答
- 最新
- 投票最多
- 评论最多
0
Hi, there are many things that you can double check to mitigate this issue, below the list of best practices:
- Check the Load Balancer Health Check: (https://docs.aws.amazon.com/AmazonECS/latest/bestpracticesguide/load-balancer-healthcheck.html) Amazon ECS waits for the load balancer to report that the container is healthy. Please check the list of parameters to speed up this process, as this could add some level of variability.
- Check the Load balancer connection draining: (https://docs.aws.amazon.com/AmazonECS/latest/bestpracticesguide/load-balancer-connection-draining.html) Amazon ECS sends a SIGTERM to the container after the load balancer reports that container draining is done. Please check the list of parameters to speed up this process, as this could add some level of variability.
- Check Container image type: depending on how big the image and in which Registry is stored you could have some variability on the time required to download the image (https://docs.aws.amazon.com/AmazonECS/latest/bestpracticesguide/container-type.html).
- Task deployment: Base on how the task deployment is configured, it could lead to longer deployment time (https://docs.aws.amazon.com/AmazonECS/latest/bestpracticesguide/service-options.html)
You can see all the best practices here https://docs.aws.amazon.com/AmazonECS/latest/bestpracticesguide/deployment.html
已回答 6 个月前
相关内容
- AWS 官方已更新 2 年前
- AWS 官方已更新 1 年前
Thanks Simone. The proper way to bypass the task failure is to adjust the health check grace period of the load balancer within the ECS service settings. I increased it to 300 seconds