In this world of micro-services and Kubernetes, it was nice to be working on a system built as a monolith running in a plain old EC2 instance. But recently, during load testing, the service we’ve been working on as been showing signs of stress — timing out, maxing out the CPU — and it’s become clear that scaling up has taken us only so far. So now might be time to rearchitect this into micro-services and move it into something like Kubernetes. This was probably always going to happen. I’m just glad that it’s happening now, rather than before it was necessary.

I guess it’s also nice that this is somehow (by now) a fairly well-trodden pattern for teams? I imagine it might have been a much more risky proposition even a few years back.

@softcow Yeah, that’s true. Certainly a few people I work with have had success with Kubernetes in previous roles (and, to be fair, so have we on other aspects of the system).