AWS ECS randomly removed target group (ec2 instance)

Question

I have an ECS service with daemon as type and today randomly the task count went from 2 (default) to 1, and haven't been able to fix it.

And the events tab is getting spammed with:
service xyz updated computedDesiredCount for taskSet ecs-svc/0123 to 1.

(daemon service xyz) updated desired count to 2.

It goes from 1 to 2 to 1 to 2 again every 10 seconds. Does anyone have an idea how to fix this? I have tried forcing new deployments, but nothing.

In the deployments tab the ACTIVE status is running one task. And the PRIMARY has desired 1, but pending and running are 0 tasks.

Answer

In my case ECS was trying to deploy an old image that was misconfigured. It was just a matter of time before my latest deploy took precedence. In my case, reviewing the logs of the containers is how I figured it out.

Also you can check the health check tab. If a health check condition cannot be met,  then it will terminate containers not passing health checks.

AWS ECS randomly removed target group (ec2 instance)

相关内容