aws resource_error ai_generated partial

service unable to place tasks: reason: capacity is unavailable at this time.

ID: aws/ecs-task-stuck-pending-capacity

Also available as: JSON · Markdown · 中文
85%Fix Rate
87%Confidence
1Evidence
2024-09-05First Seen

Version Compatibility

VersionStatusIntroducedDeprecatedNotes
ecs-2024-06 active — — —
fargate-1.4.0 active — — —
ecs-cli-2.0.0 active — — —

Root Cause

ECS cluster lacks sufficient EC2 or Fargate capacity to launch the task, often due to instance type unavailability in the AZ.

generic

中文

ECS 集群缺乏足够的 EC2 或 Fargate 容量来启动任务，通常由于可用区中的实例类型不可用。

Official Documentation

https://docs.aws.amazon.com/AmazonECS/latest/developerguide/service-event-messages.html

Workarounds

  1. 85% success Change the task placement strategy to spread across multiple AZs: `aws ecs update-service --cluster my-cluster --service my-service --placement-strategy type=spread,field=attribute:ecs.availability-zone` 
    Change the task placement strategy to spread across multiple AZs: `aws ecs update-service --cluster my-cluster --service my-service --placement-strategy type=spread,field=attribute:ecs.availability-zone`
  2. 75% success Use a different instance type or Fargate platform version: update the task definition to use `FARGATE` or a different CPU/memory combination. 
    Use a different instance type or Fargate platform version: update the task definition to use `FARGATE` or a different CPU/memory combination.

中文步骤

  1. 更改任务放置策略以跨多个可用区分布：`aws ecs update-service --cluster my-cluster --service my-service --placement-strategy type=spread,field=attribute:ecs.availability-zone`
  2. 使用不同的实例类型或 Fargate 平台版本：更新任务定义以使用 `FARGATE` 或不同的 CPU/内存组合。

Dead Ends

Common approaches that don't work:

  1. 90% fail

    Restarting the ECS service doesn't free capacity; the underlying resource shortage persists.

  2. 80% fail

    Increasing task count makes the problem worse by requesting more unavailable capacity.