Amazon Elastic Container Service improves accuracy of Service Load Balancing
Amazon Elastic Container Service (Amazon ECS) has improved the accuracy of Elastic Load Balancing (ELB) for Amazon ECS services. Load balancing on Amazon ECS now more accurately routes traffic to running tasks as tasks will be deregistered from the ELB before they enter a stopped state. Furthermore, with these improvements, Amazon ECS also helps your services running on the Fargate Spot capacity provider be more resilient to Spot termination notices.
Amazon Elastic Container Service (Amazon ECS) has improved the accuracy of Elastic Load Balancing (ELB) for Amazon ECS services. Load balancing on Amazon ECS now more accurately routes traffic to running tasks as tasks will be deregistered from the ELB before they enter a stopped state. Furthermore, with these improvements, Amazon ECS also helps your services running on the Fargate Spot capacity provider be more resilient to Spot termination notices.
Amazon ECS can be configured to integrate with ELB to distribute traffic evenly across the tasks in your service. Amazon ECS has improved its ability to properly de-register tasks from the load balancer’s target group, thereby reducing the potential for transient errors should traffic be routed to stopped tasks. With this improvement a task is now deregistered from the ELB before the task enters a stopped state, leading to both improved accuracy for traffic routing and a material reduction in traffic routing errors experienced by Amazon ECS customers. Additionally, Amazon ECS will now deregister your task running on Fargate Spot, if it receives a spot termination notice, before issuing a SIGTERM message to inform the task that it needs to stop. This improvement helps you manage spot interruption of tasks running on the Fargate Spot capacity provider more safely.
To learn more about load balancing on Amazon ECS, visit the Amazon ECS Documentation. This improvement has been made available across all AWS Regions.