> Your goal, when launching a product, is to build a product that solves a problem for your users. Not build the fanciest deployment pipelines, or multi-zone, m

> Your goal, when launching a product, is to build a product that solves a problem for your users. Not build the fanciest deployment pipelines, or multi-zone, multi-region, multi-cloud Nuclear Winter proof high availability setup. I agree, but context is key: if you're bootstrapping a start up, you don't need these things. You need to prove your product, then you scale. But automation != scale. Having a process that streamlines your delivery process, regardless of scale, can be helpful. I've screwed up enough single-box deployments to learn that less. Stepping back: our industry is pretty horrible about creating tools that can start small and scale up. I like where CockroachDB is going for that reason (just as an example). It would be great to start with a single database, and have a clear path to scale it horizontally across multiple nodes and data centers. Kubes might get there... I'm not sure how focused they are on making small things work well, though... any examples of that?