Introducing new alerts to help users detect and react to blocked job queues in AWS Batch

The article discusses new features in AWS Batch that help users detect and react to blocked job queues. AWS Batch allows running batch workloads on Amazon ECS and EKS, and sometimes a job at the head of the queue can block other jobs from running due to issues like misconfiguration or lack of resources.

Specifically, the article covers:

• Common causes of blocked job queues, including insufficient capacity, compute environment limitations, job resource requirements, service role permissions, and invalid compute environments.
• New CloudWatch Events notifications that Batch sends when it detects a blocked job queue, including the specific reason for the blockage.
• Using Amazon EventBridge to define automated actions (like terminating the job) when receiving these notifications.
• A new jobStateTimeLimitActions parameter that allows automatically cancelling a stuck job after a specified time.
• A table mapping different blocked job queue scenarios to the corresponding CloudWatch event messages and parameters for jobStateTimeLimitActions.