I run an app with 1m+ DAU. For me, the server tech stack did not matter. What did matter was the ability to iterate quickly and identify bottlenecks. Although it was still critical, the server was not a massive component of the overall business, which may have given me better mileage than other users with a sub-par tech stack.

I originally started the server component using procedural PHP on $5/month shared hosting, which ground to a halt after receiving local news exposure. As users grew, I slowly built up the server stack to be more robust, tested, etc. For the most part, my users were okay with these growing pains (certain parts of the service had downtime as it expanded and grew) and they did not impact the overall success of the app.

When starting a new product, my biggest focus would be to put together the stack in a way which allows for the ability to expand, pivot, and change to account for unforeseen events. I feel that in a lot of cases, too much time is wasted debating tech stack factors which ultimately do not matter.