when to optimize for latency versus throughput? dont worry, its not as complicated as it sounds.

Hey everyone, I've been thinking a lot about system design trade-offs lately, and especially about latency versus throughput. For some services, it's super important to make sure individual requests are fast, you know, like for interactive user interfaces. But for other services, the main thing is handling as many requests as possible per second, which is key for things like batch processing. So, what are your usual approaches or quick tips for deciding which metric to focus on when you're first designing something? Are there common ways of doing things or particular technologies that really push you towards one or the other?