The thing is there is no hivemind best practice. It depends on a lot of factors such as team size, project duration, technology stacks, etc... As you may already know being able to spot pros and cons ahead of time and avoiding pitfalls for a specific scenario just comes from experience. So if anyone is uncertain about how to proceed, the only way you can become more comfortable in decisions is to try all the things and learn the nuances of different approaches. Eventually it becomes much more natural and requires less effort.