When is serverless too much, compared to just using a small, always-on VM for background stuff that doesnt happen often?

Hey everyone, I've been trying to figure out the best way to handle a new internal service. It's for background jobs that don't happen often and aren't super important, like daily data exports for a few teams. My first idea was serverless functions, like Lambda or Cloud Functions. That feels like the smart, new way to do things, and I wouldn't have to deal with servers at all. But these jobs are really infrequent, and they'll probably run for a very short time. So, I'm starting to think: what if I just set up a tiny VM, like a t3.nano or e2-micro? It would just sit there mostly idle, running a cron job. That might actually be easier and cheaper in the long run. We don't need huge scalability or sudden bursts of power. Am I making too big a deal out of the cost of an idle small VM, thinking serverless is always better here? Or am I missing some tricky parts of using a VM that serverless just takes care of? What do you all think about choosing between these for tasks that don't happen often?