when is eventual consistency good enough, really?

hey everyone, i've been thinking a lot about consistency models lately, especially in distributed systems. we often strive for strong consistency, but there are times when eventual consistency seems perfectly acceptable, or even preferable, to avoid performance bottlenecks. what are your thoughts on when eventual consistency is 'good enough'? are there specific use cases or scenarios where you consciously choose it, and what are the trade-offs you consider