How do I troubleshoot schedule issues for Amazon ECS tasks that run on EC2 instances?

Resolution

If Amazon ECS can't place your task, then the task remains in the PROVISIONING state. You receive the following error in your service event messages:

"service ServiceName was unable to place a task because no container instance met all of its requirements."

To troubleshoot this issue, check your service event messages for this error. Then, complete the troubleshooting steps based on the error reason.

No container instances were found in your cluster

If you don't register EC2 instances to the cluster, then you receive an error. Check for scaling or instance launch issues with the capacity provider or Amazon EC2 Auto Scaling group that you configured.

If the instance launched, then check why it isn't connected to the cluster.

The closest matching (container-instance abcdexyz) has insufficient CPU/memory/GPU units available

This issue occurs when the available container instances don't have the CPU, memory, or GPU that your task requests. To resolve this issue, take the following actions:

• Reduce the resources that you allocated to the tasks in the task definition if it's more than the container instances can manage.
• Increase the number of container instances in the cluster.
• If you can't increase the number of container instances, then use larger container instances with more resources.
• Check whether you can reduce the number of task replicas.

To further troubleshoot issues with CPU, see How do I resolve "the closest matching container-instance container-instance-id has insufficient CPU units available" error in Amazon ECS?

The closest matching (container-instance abcdexyz) already uses a port required by your task

This issue occurs when the container instance already uses the required port. You can't run multiple tasks on the same port. To resolve this issue, add more container instances to your cluster. It's a best practice to automatically scale your container instances. Create a capacity provider for your cluster. Or, to allow your tasks to use any available port on a container instance, use dynamic port mapping.

The closest matching (container-instance abcdexyz) misses an attribute required by your task

This issue occurs when your container instance is missing a required attribute defined in the task definition. To check for missing attributes, run the check-attributes Amazon ECS command line interface (ECS CLI) command:

ecs-cli check-attributes --container-instances container-instance --cluster cluster-name --region region-code --task-def task-def-name

Note: Replace container-instance with your container instance name, cluster-name with your cluster name, region-code with your AWS Region, and task-def-name with your task definition.

The output of the preceding command shows a table with the missing required attributes. For information about how to troubleshoot missing attributes, see Checking for missing attributes and debugging reason attribute errors on the GitHub website.

The closest matching (container-instance abcdexyz) doesn't have the agent connected

This issue occurs when the Amazon ECS agent disconnects. To troubleshoot this issue, see the following articles:

• If you use Amazon Linux 1 (AL1), then see Why are my Amazon ECS container instances with Amazon Linux 1 AMIs disconnected?
• If you use Amazon Linux 2 (AL2) or Amazon Linux 2023 (AL2023), then see How do I troubleshoot a disconnected Amazon ECS agent?

The closest matching (container-instance abcdexyz) encountered error "MemberOf placement constraint unsatisfied."

If you use placement constraints to place tasks in specific instances, then make sure that the container instances support the selected attributes or constraints.

For more information, see How do I troubleshoot issues with Amazon ECS task placement constraints?

Related information

Amazon ECS troubleshooting