One thing I've noticed with software projects is that getting a prototype working is often the easy part. The difficult decisions tend to appear later: architecture, testing, monitoring, security, deployment, technical debt, and making sure the system can handle real users.
I'm curious how other developers here approach that transition. At what point do you start treating a project as “production software” rather than an experiment?
For example, do you prioritize observability and automated testing early, or wait until the product has proven demand? And how much architecture do you consider necessary before the first version goes live?
I've been looking at how teams such as Software Developers Pro approach this problem, and their emphasis on production engineering rather than stopping at the prototype stage got me thinking about it.
For developers who have taken projects from prototype to production, what was the biggest issue you didn't anticipate?