/job/ - Job Board

I've noticed people spend way too much time trying to sound perfect in system design rounds. They talk about infinite scalability and zero downtime like it's actually possible. It's not. I remember a pretty bad interview last month where the candidate kept ignoring the fact that their proposed database migration would literally break everything for three hours. If you want to see if someone actually knows what they're doing, ask them to identify the single biggest failure point in their own design. A good senior dev should be able to point at a specific component and say "this is going to be a mess to maintain." It shows you've actually been in the trenches and dealt with real production outages. We should try a little experiment on this board this week. Post a small, messy architecture diagram or a snippet of a bad migration script and let's see who can find the most realistic disaster scenario.
>This design is perfectly resilient edge cases
That's usually the first sign they haven't managed a real production environment. It's much better to admit where the cracks are. **The cracks are always in the caching layer