That is wild!
A little update for all of you interested. I allocated the additional resources to the VM and will post some updated graphs once they update with the new configurations.
For those who are like me and like looking at graphs here are some prior to the upgrade.
CPU - 48 hours
Memory - 48 hours
Network - 48 hours
It can definitely run on a Pi instance. The storage medium is going to be important as it will need to house a PostgreSQL database. Would be fine for a few users but not sure how many it would be able to handle concurrently.
How quick does this instance load for you compared to other sites you visit? I'm curious to how well it peers with Baghdad!
Was the image above generate by AI? If so what software are you leveraging for this?
Sure lets see how far we can get this server going!
A couple hundred gigs worth of ram and around ~100 cores. I'm not sure I would want to bring it to that size without thinking first of a plan to properly maintain, manage and support.
It probably doesn't need that much but I'm curious in seeing how much It can handle until it can't I have more resources available but it would make more sense after this size to start separating the different components of Lemmy into their own individual instances across multiple servers instead of trying to scale it vertically. Might require a little more time than I have but I'll see what I can manage!
No Arista? It's been really great from my perspective. basically identical to the Cisco IOS!
Sweet! Welcome!
The only part I don't have graphed yet is the disk IOs. I'm going to need to invest a little more time to get that metric going captured.
Say you through 1000s of active users to this instance.. what would cause the bottle neck first? CPU, memory, network, disk? I'm thinking probably disk due to database optimization that need to be reworked on lemmy afterwards CPU and then memory.