I'm probably wrong about my assumptions, but presume you are open to any kind of constructive feedback, so here it goes... Maybe you're overkilling with the in

I'm probably wrong about my assumptions, but presume you are open to any kind of constructive feedback, so here it goes... Maybe you're overkilling with the infra stack. I would simplify until having a mature product, especially if I'm bootstrapping, which I think is your case. Right now, you're still a bit far from MVP, from my point of view. Those $200 can probably be reduced by 50%-75% if you compromise on stuff only important to non-alpha services (i.e. 99.99% availability). A single EC2 box should be enough. Maybe look into Postgres or another FOSS instead of BigQuery. These $100-$150 savings per week can go into promoting your service, getting as much attention as possible to maximize feedback. Good luck!