[+][FCB] Cats effects was giving warning on starvation, increase throughput...

[+][FCB] Cats effects was giving warning on starvation, increase throughput performance for delta messages via chunks

+ 1st it was attempted to make it fully granular, but...it does not behave well. Keep it in a branch and check another time.
+ It feels like every message on the redis stream opens a whole IO thing and the controller was eating all the memory very fast.
+ By grouping/chunking the problem gets put under certain control.
+ Testing showed to handle 20K delta items with grace, not instant, but gracefully. Felt like a good balance for redis and the controller.
+ Smaller chunks was giving starvation warnings, larger it's quite impactful on memory due to deser.
+ Maybe it's time to do binary streams, specially in this particular area of the code.
+ In any case, 500 chunks handles 20K properly. Still not the desired performance, but also, IPFS cluster is quite a bottleneck here.

It's kind of a tricky problem. Newcomers, specially big members with over >100K files might find it unstable to put their shared folder online and this would be their very first experience with the software...
But also...what are the chances that such a big member would it find unacceptable to have to chunk their file population?
Still, no compromise, aim for handling millions of files with no problem.

What's left then is:
+ Checking why does the CPU starvation happens. In >100K cases, the problem will still be there.
+ Binary streams. RPC to the rescue? This would need some serious metrics.
+ At some point, the file tree resolution will also act as a bottleneck. For 15K it's 2MBs in Redis and quite some memory pressure in controller's backend. Some solution will be needed here at some point.