With the popularity of containers and container-based architecture, it can be hard to keep up with new container technologies and pick out the useful ones from a sea of competition. It’s important to carefully consider technology choices lest one adopt technology they don’t really need for minimal gain, and fail to KISS.
Bottlerocket was released for public preview by Amazon Web Services (AWS) in early March 2020, and for general availability on August 31, 2020. Its purpose is to lower management overhead and costs and improve security, applying many of the principles behind Docker to the instances we run Docker atop of. It does this through locked-down access and bare-minimal software on instances running Bottlerocket, greatly reducing the attack surface.
- No additional cost. Standard EC2 rates apply.
- “Atomic updating.” Updates are applied in a single step, rather than by package, with the goal of reducing management overhead and making rollbacks easier.
- Supports Open Container Initiative Image Format images.
- Backed by Amazon. A huge company’s support improves Bottlerocket’s odds of not going the way of Container-optimized OS and Container Linux.
- AWS ecosystem focused. Bottlerocket is built by Amazon, for use within AWS.
- No SSH, no shell. Access is intended to be through orchestration tools, e.g., EKS, not directly.
- Requires containerized architecture, e.g., microservices running in a Kubernetes cluster.
- Relatively new and immature.
Most of these cons, such as locked-down instance access and container focus, are by design, and not strictly disadvantageous. Given the lack of additional costs, if you’re already ingrained in the AWS ecosystem, there’s no reason not to give it a shot.