When does a read-heavy system really need CQRS? I mean, when is it worth the extra complexity?

Hey everyone, I'm dealing with a system that's suddenly getting a ton more reads but not many writes, and the reads are starting to cause some serious slowdowns. I've been thinking about trying Command Query Responsibility Segregation, or CQRS, to split up how we handle reads and writes. Has anyone actually used CQRS for situations where you have way more reads than writes? How did it go for you? What were the biggest downsides you ran into, and when do you think it's actually worth all the extra trouble compared to easier ways to handle read scaling, like read replicas or just adding more cache?