Amazon Elastic Container Service (Amazon ECS) vs. Apache Mesos

Amazon Elastic Container Service (Amazon ECS) is well suited where you need the ease of managing the clusters by letting AWS do the stuff for you. Obviously, whenever you want to run the docker based workloads, it is always better to go for either AWS ECS or AWS EKS. If you are interested in staying at AWS only and don't want to be cloud-agnostic, then go for AWS ECS instead of AWS EKS. AWS ECS is cheaper than AWS EKS and also more managed by AWS and better integrated with other AWS services. If you want to run those workloads as serverless, then AWS ECS Fargate is the best option to go with. If you already have a Kubernetes based setup that you want to migrate to AWS, then go for AWS EKS instead of AWS ECS.

Incentivized

Mesos is really great when you have a big datacenter with many different applications and use cases. It will help you to optimize the resource usage, being a centralized API for your infrastructure. It will not suit well for small companies that just need to deploy a web app. In this case, I would recommend something smaller.

Incentivized

• Well Integrated - As with the majority of AWS services, ECS works will with any other AWS product (Route 53, CloudWatch, IAM, etc).
• Easy to get started with - It is easy to get started building just about anything in AWS and using ECS is no exception to this rule. Be careful though -- AWS lets you do/build anything in any way you could think of and allowing yourself to shoot yourself in the foot is no exception.

Incentivized

• Mesos may have many frameworks. If you have Mesos installed on your servers, you may use it for many kinds of tasks. Today we're running only web applications but the idea is to install a different framework for big data soon.
• There is a good community growing around it.

Incentivized

• The user interface sometimes seem to be confusing and cumbersome. It can be improved so that people can understand clearly which section to go for which functionality.
• When a container fails, the error logs are not readily available on the ECS console. If it can be provided it would be easier to debug from there itself instead of going to our log manager.
• Sometimes the old EC2 containers become stale and need to be restarted manually. There should be a notification for such scenarios. We have mostly been finding it out on our own and then fixing it by manually restarting EC2 instances.
• If this could be proactively monitored and notified, it would be great.

Incentivized

• Unreliable deployments that would fail for no good reason. Sometimes our Docker container would be "restarting" forever because Mesos thought it didn't have enough resources to start the container.
• Impossibly slow UI. Built in React under the hood with a lot of bloatware backed in, so loading the Mesos UI on a slow internet connection was painful.
• No real logging solution - it would stream "console.log()" output to the UI, but searching for logs wasn't really possible without downloading a huge file.
• No built-in support for redeploying containers from a CI. We had to create a service whose whole job was to expose an HTTP endpoint that restarted a container, and then made Circle CI ping the endpoint whenever we wanted to redeploy.

Incentivized

Aside from some ECS-specific terms to learn at first, learning & starting to use ECS is relatively straightforward. AWS docs on the topic are also of high quality, with sound & relevant examples to follow. Troubleshooting container issues is also a breeze thanks to CloudWatch integration & helpful error messages on the AWS console.

Incentivized

Support is relatively good, although the documentation sometimes is lacking, as well as outdated in our experience, especially when we initiated the process of using this service. But once we found how to assemble things, we haven't really required support from anyone at AWS, the service works without problems so we haven't had the need to contact support, which speaks well of how ECS is built.

Incentivized

No real support channel, the Mesos GitHub issues list was the only one we found and it wasn't particularly helpful.

Incentivized

AWS Elastic Beanstalk is a good beginner level orchestration service but lacks container management and scaling capabilities. EC2 is again not a Managed cloud service. It is like just renting a computer on cloud and then managing it on our own. Compared to these ECS is a comprehensive solution that provides management, scaling, containerization and other service connectivity out of the box.

Incentivized

Kubernetes is by far the best choice. More reliable and better developer experience. Mesos is prone to sporadic failures and not really designed to handle CI/CD-based deployments. Docker Cloud once shut down our entire cluster for "upgrades" without giving us any warning.

Incentivized

• We run 8 web applications (demo instances) on a single machine. At a particular time, no more than 3 applications run simultaneously. So, we keep only required containers up. This helps us to provision small EC2 machines without compromising performance.
• Overall Amazon ECS helps to have less number of dedicated machines as more than one solution can be deployed on a single instance. This reduces costs a lot.

Incentivized

• Developers needed to spend hours waiting to verify that their supposedly automated CD deploys actually went through.
• Lots of late night and early morning calls that Mesos deploys are hanging again.
• 10-20 hours spent building a custom service just to integrate Mesos with a CD provider.

Incentivized