when should you start thinking about data lakes

hey everyone, seeing a lot of talk about microservices, queues, and scaling lately. it got me thinking about data storage. we're currently running a pretty standard relational database for our main application, and it's working fine. but as we collect more diverse data, like logs, user behavior, and external APIs, i'm wondering when it makes sense to start considering a data lake. is it purely about volume, or are there other triggers, like needing to do complex analytics across different data types? what are your experiences