Kuberentes Hosting Services

Overview

When looking for a hosting platform for Kubernetes, I wanted to find a platform which was easy to use, had a good developer experience, and that was cost-effective. Easy to use is somewhat subjective and certainly depends on familiarity with the platform, domain knowledge, and other factors. Therefor, I’ll try to be as objective as possible when evaluating the platforms looking at Developer Experience and Cost Effectiveness.

For others, there could be other dimensions which are more important. For example, if you’re looking to meet certain compliance requirements, you might want to look at the security and compliance features of the platform and rate them accordingly.

For me and my project, these are not yet significant concerns.

Hosting Platform Options

An AI Assisted search via OpenAI’s ChatGPT¹ for Kubernetes hosting platforms yields the following results:

Query:

create a markdown table which includes the following hosting providers:
AWS
Google Cloud
DigitalOcean
Azure

use the following columns so that each option could be evaluated:
- developer experience
- cost effectiveness

developer experience should include productivity, impact, satisfaction
cost effectiveness should include components and pricing for those components

Validating the Findings

Cost Effectiveness

The following are specifications for a development environment. The goal is to have a non-high availablilty Kubernetes cluster with 2 worker nodes intended for a development environment. The cluster should have a managed control plane and managed worker nodes, and should have object storage and load balancing. The cluster should also have a managed Kafka instance.

Pricing has been calculated generally using two worker nodes, and the cheapest option for the managed control plane.

Monthly Pricing (as of November 2023²):

Developer Experience

GitHub mentions Developer Experience (DevEx)⁷ as productivity, impact, and satisfaction. My thought is to document my experience so that other’s can evaluate the platforms for themselves.

Given the pricing schedule above, it’s not currently feasible for me to fully evaluate all the platforms at the same time. Instead, I’ll focus on the most cost-effective one, DigitalOcean. If given the opportunity and necessity, I’ll evaluate the other platforms in the future.

In a follow-up article, I’ll report my observations and experience. For now, I’ll leave this as a placeholder.

Thanks for reading!