When I say something about load balancing, may people think about working with the existing load balancers from our native cloud providers. This is some sort of off the shelf component that will save your works. Sometimes, you need to separate your own app into multiple load balancers to make sure your app is running correctly.
Let me give you an example, you have to store a million of movies stored in your storage pool. When a user access a movie file, can you imagine how will it be possible to search across a movie out of a million movies storage. When you use the cache, it will up in the cache but multiple region has multiple requirements and limitations. How to balance the load for localizing content without hitting the main server every time.
Another example is I have a server firm about shortening a lot of URLs through my website. I would like to balance the load according to the index or the domain name without hitting everything on my main server. Then the load balancing on my sibling servers is a must. When the visitor hit the client website, then the load balancing architecture on my application can handle the short URL and translate into Long URL for the sake of ease.
This is a bit challenging due to the workload of the devs and your server architecture. I am sure that this is nothing less profit than expending the whole lot of your servers to meet your needs.