One of the problems with the Web 2.0 monolith pattern (server + web client) is that it makes community-hosting of services seem much more resource-intensive than it needs to be.
How much of the server resources used by a standard Matrix homeserver install, for example, is used by the Element web client? How much cheaper would it be to run a headless Matrix server, and use native Matrix apps on whatever devices you use?
EDIT: Or a headless AP server + native apps?
Hardware requirements listed at https://docs.element.io/latest/element-server-suite-classic/requirements-and-recommendations/#hardware
> Element recommends a minimum of 8 vCPUs/CPUs and 32GB ram for the host(s) running synapse pods.
Add a couple GB for microk8s ;)
This seem to be a common assumption. You got some numbers to back that up? Generated from a repeatable method?
> nothing's stopping you from just blocking the routes to serving the frontend files in the webserver
Nothing's stopping me from mowing my lawn with a pair of nail scissors either. But the purpose of automation and software packaging is the do jobs for us, not give us extra jobs.
(1/2)
@evan Instead of settling the question by suggesting a way to objectively measure the resource usage in question, @janw just took offence and blocked me. So I'm just going to assume it was bad faith engagement from the get-go 🤷♂️
If anyone knows of repeatable methods for measuring delivery of web apps, as a proportion of server resource usage in server + web client monoliths, sing out!
I'm still very curious about this.
(2/2)
It runs ob the client
Electron is a UI library
