Gregory's Server@uastronomer @george Arguably a t3.medium is overpowered for the current workload (I'm averaging like 5-10% CPU) but (a) deployments and frontend rebuilds are CPU-intensive and (b) I'd rather have the headroom to deal with surges in activity + ensure decent uptime. But we'll see how it goes!
@wogan @george That's actually not too bad!
I'm hoping for way lower costs per user though, since I'm pushing small private instances.
mastodon.monoceros.co.za is single-user, so far, and is doing okay on a t3.micro, with no ELB and a postgres RDS instance. Coping well so far, based on watching CPU burst credit usage. We'll see how it copes over time though.
Meanwhile, work continues on automating as much of the New Instance Creation process as possible. I als
@wogan @george That's actually not too bad!
I'm hoping for way lower costs per user though, since I'm pushing small private instances.
mastodon.monoceros.co.za is single-user, so far, and is doing okay on a t3.micro, with no ELB and a postgres RDS instance. Coping well so far, based on watching CPU burst credit usage. We'll see how it copes over time though.