Platform engineering encompasses a broad domain with continually evolving frontiers that require overcoming new challenges to level up the solutions offered to engineers. However, this evolution can lead to situations where previously-made assumptions may no longer be applicable to new solutions.
Our team recently encountered this scenario when using AWS Stacksets across multi-account/region rollouts of our platform, which highlighted some interesting dependency challenges that we only discovered at the 11th hour of our project. The new world introduced edge cases that we didn’t anticipate, and we learned that it’s crucial to identify unknowns early and be honest about them. It’s essential to deliver a first step that cuts across those unknowns as soon as possible and iterate from there, even if that means doing something hacky. If a hack doesn’t work, it’s better to discard it than a polished solution that would never work.
Ultimately, our lesson was to de-risk early and embrace a growth mindset that values learning from failures and leveraging them to improve.