Category: Amazon Elastic Container Service

Amazon ECS Managed Instances provides built-in GPU health monitoring and automatic repair powered by NVIDIA DCGM. This post shows how GPU auto-repair detects critical GPU hardware failures and replaces impaired instances with no code changes, and how to observe and safely validate the workflow using DCGM fault injection on a dual-tier inference pipeline.

Amazon ECS now publishes high-resolution 20-second CPU and memory utilization metrics for target tracking, so service auto scaling detects and responds to load changes faster. This post shares AWS benchmarking results across AWS Fargate and ECS Managed Instances and shows how to turn the feature on.

Ramp runs GPU-powered AI inference continuously on Amazon ECS. This post walks through how Ramp’s infrastructure team moved those GPU workloads onto Amazon ECS Managed Instances: the architecture pattern, Terraform implementation, and the lessons learned migrating 50 to 60 EC2 instances.

Learn how to encrypt traffic between Amazon ECS workloads using two native approaches: VPC encryption controls for network-layer encryption through the AWS Nitro System, and Service Connect TLS for application-layer encryption. A hands-on walkthrough shows how to activate encryption on AWS Fargate and verify it in VPC Flow Logs.

Running an OpenTelemetry collector as a sidecar in every Amazon ECS task does not scale across a multi-account estate, and it cannot run at all on Windows. Learn how to replace per-task sidecars with a single centralized ADOT gateway that ingests OTLP from workloads across accounts and exports traces to AWS X-Ray and metrics and logs to Amazon CloudWatch.

Amazon ECS Express Mode gives you load balancing, scaling, logging, and networking out of the box. Learn how to extend an Express Mode service beyond its defaults with three hands-on examples: turning on ECS Exec, adding a scheduled scaling policy, and adding a FireLens log router as a sidecar container.

Deploy a single CloudFormation stack that builds a GPU batch inference pipeline on Amazon ECS Managed Instances. It uses Amazon SQS for job buffering and Application Auto Scaling to scale to zero when idle, so you pay only for active inference time.

In this post, we explain how zone-aware routing works and walk you through setting up a multi-AZ ECS cluster to see it in action.

Starting on June 10th, 2026, Bitnami container images will no longer be available on Amazon ECR Public Gallery. If you currently pull Bitnami images directly from ECR Public in your workloads, you need to take action before this date to avoid service disruption. In this post, we walk you through how to determine if you’re affected, how to mirror the images you need to your own private registry, and best practices for protecting your workloads from future upstream changes.

In this post, we walk through how linear and canary strategies work in Amazon ECS, how to configure each, and how to set up automatic rollbacks with CloudWatch alarms.