I would assess the root causes of scalability issues, prioritize critical features, and adjust timelines. I’d communicate transparently with stakeholders about the challenges, propose solutions, and allocate resources to ensure timely delivery without compromising quality.

Software design is a great long-term approach to solving scalability challenges (as opposed to throwing more hardware at the problem, which is too common). If you're designing a prototype or even MVP it's questionable how scalable the first incarnation of your design needs to be, but let's assume there is a legitimately rationalized near-term scalability challenge to overcome. I suggest building a client-side test harness to synthesize user transactions at scale (perhaps even beyond the target scale). Overwhelm your application and analyze the tipping point with introspection, profiling, APM, etc. Re-design the mechanisms which bottleneck your application. Break out slow parts into new services/systems purpose-built lighter with Rust or C.