There are so many interesting problems to solve with regards to @pixelfed and scaling/performance
One example is how I'm debating weather to keep a full redis sorted set of followers for each account or just a hundred and use db for pagination (when appropriate) to reduce RAM and make more room for other features.
It's basically a game of whack a CRUD where you're hitting bottlenecks with cache but then cache become bottlenecks 😅
@dansup @pixelfed initially just let your DB handle caching for those situations, it is made for it. Of course that does require proper queries and decent indexes. PostgreSQL is great for that reason as it has the analysis/explains where one can update things. Adding a redis etc, is side cache, allowing your DB to cache more can be more than enough. Less moving parts, less cache inconsistencies.