I'm exploring the idea of creating a container orchestrator with a built-in gateway and service mesh that can connect multiple VPSs to form a cluster. The goal is for each VPS to manage external requests for services within the cluster, either routing those to local containers or to containers on other nodes through the mesh service. I'm envisioning this all working from a single binary on a small VPS that still has capacity for a few additional containers. I realize that using Docker or Kubernetes isn't viable due to their resource demands, but could a more hands-on approach with the underlying technologies they use be the answer? This setup would potentially cut costs on AWS by eliminating the need for ALBs and VPCs, and with the built-in gateway, it would be ready to serve requests immediately. So, is this actually possible?
3 Answers
Honestly, it sounds pretty complicated to set up without using existing solutions. You might run into issues with scaling and managing workloads unless you have significant experience in orchestration. I'd suggest building on what already exists rather than going it alone. That route could save you a lot of headaches!
I think it’s a stretch to say you can create something stable that competes with Kubernetes. Without substantial field experience, the idea may not hold up. Better to learn the ropes first and then innovate!
What you're describing sounds pretty ambitious! While a lightweight orchestrator is a cool idea, keep in mind that Kubernetes is designed for a reason. Trying to recreate all its functionalities without the overhead might just lead to a lot of tech debt in the long run. It’s awesome to think outside the box, but you might end up spending more time fixing issues than enjoying the benefits. Just something to consider!
Exactly! The principles behind Kubernetes and similar tools have been refined over years. While a custom solution might be feasible, managing all the complexities can be a real headache.
Yeah, I agree. It's tempting to innovate, but sometimes leveraging established tools is the smarter choice. It can save time and more importantly, reduce possible bugs.