Most of our apps, on a single t3a.nano (about $3/mo) can handle about 250 req / sec in stress tests. In sluggish python no less. People don't seem to understand modern compute speeds.