Node.js API development woes

hey everyone, i just launched my tiny SaaS app, and while things are fine in dev, my production API development endpoints are acting like a moody teenager. seriously, one minute they're fine, the next they're throwing a tantrum. it's driving me nuts trying to figure out what's going on.

the main issues i'm seeing are:

• randomly slow responses for certain api development endpoints. it's not consistent, sometimes a specific endpoint is lightning fast, other times it takes ages.
• sometimes gets 500s even when i haven't touched the code or deployed anything new. it's like there's a ghost in the machine.
• database connection issues (using Mongo Atlas) seem to crop up intermittently only in prod, never local. my local dev environment is buttery smooth with the same database connection string.

i've tried a bunch of stuff already, hoping to squash these bugs:

• checked server logs (AWS EC2), but nothing immediately obvious screams 'this is the problem'. just a bunch of standard access logs and the occasional unhandled promise rejection, but not correlating with the 500s.
• optimized database queries as much as possible, added indexes, etc. i thought maybe it was slow queries, but even simple ones sometimes hang.
• added more memory/CPU to the EC2 instance, thinking it was a resource bottleneck. no real change, still seeing the same intermittent issues.
• tried different Node.js versions, hopping between LTS versions just in case.
• used pm2 for process management, thinking it was a crash issue and pm2 would restart, but the problems persist.

so, my question to you lovely folks:

• is there some common pitfall with Node.js/Express API development for small to medium scale apps that i'm totally missing? something everyone 'just knows'?
• are there specific tools or monitoring strategies for identifying these intermittent backend issues that don't show up clearly in standard logs? i'm open to anything.
• should i consider a different framework for better stability or just optimize my current setup further? i'm invested in Node.js but stability is key.

anyone faced this before with their backend?