Amazon Elastic Container Service (Amazon ECS) Alternatives

We evaluated several managed Kubernetes options before going with EKS about 2 years ago and it’s been a great choice for our container orchestration needs. The Kubernetes API access is fully conformant, which was a non-negotiable for us. On the container runtime side, the shift to containerd has been seamless. We haven't run into compatibility issues, and the performance has been stable across our production workloads. EKS managed the Docker to containerd transition better than I expected.

Google Kubernetes Engines is a highly scalable managed kubernetes platform that simplifies the application deployments, management and operation especially for high traffic environments where I am working (E-commerce website)

I used Fargate to run containerized workloads without managing EC2 nodes. This removes a significant operational burden. Scaling, patching, and node lifecycle management are no longer a concern The tradeoff is a higher per-task cost compared to normal nodes

My overall experience with the product has been very positive, as I migrated an internal environment that was on VMs to AKS OOP, and the improvement was noticeable in terms of cost and performance.

The platform provides a reasonably stable K8s-based orchestration environment with solid integration to other Huawei Cloud services (networking, storage, and security). CCE aligns well with standard K8s workflows, however the documentation quality of these services is inconsistent and troubleshooting guidance can be limited. Some advanced configurations require more manual intervention than expected, for example, checking how many and which deployments are using a specific PV or PVC storage. Our company is also actively investigating performance inconsistencies in our production deployments. While the orchestration and autoscaling configurations are implemented correctly based on official documentation, the platform has not consistently delivered the expected website performance improvements under scaling events. This has required additional performance analysis and tuning to determine which technological layer the bottleneck originates from.

So, SUSE Rancher RKE2 (Rancher Kubernetes Environment for Government), as a purpose-built platform for regulated and security-sensitive environment, the opinionated hardened defaults fit well within our operations. Overall, RKE2 and Rancher provide one of the most stable and repeatable Kubernetes experiences available for government customers. Rancher simplifies our operational overhead of deploying, upgrading, securing, and managing Kubernetes at scale. The platform has consistently demonstrated reliability and predictability.

Portainer does the job it's built for, but it also isn't the best container management/orchestration platform out there. It handles building and deploying containers into your environment and supports deployment to remote environments, which is very convenient at scale. However, Portainer's UI/UX is clunky, and I find it easier to deploy out of CLI and YAML files instead of using Portainer due to Portainer's own unique methods of handling deployments.

My experience with Google Cloud Run has been very positive, offering simple deployments, automatic scaling and reliable performance for a containerized React application. Provides reliable scaling and reduces operational effort to manage the infrastructure.