Open turt2live opened 5 years ago
https://github.com/matrix-org/matrix.org/issues/298 already existed but this already has 👍 so I'll close that instead.
Some info here https://matrix.to/#/!QtykxKocfZaZOUrTwp:matrix.org/$15448853122944541SfWfb:matrix.org
oops, sorry. I'm the worst.
@omame do you think this is something we can do?
It's tricky as they tend to change, especially in recent times. And the situation will only get more confusing in the future. We are going to revamp the hardware and apply a number of changes as we increase the reliability of the service.
The outcome of this is that the information we publish will likely be outdated shortly thereafter.
If the goal is to get a rough approximation, then I believe there are some resource calculators out there that spit out the resources needed to support a deployment for x users. Give its size, the matrix.org homeserver is more of an outlier than a common use case, so it wouldn't be very useful as a reference.
The core Synapse machines don't change size very often though? Honestly, even if the specs are updated annually then that would be fine - people are mostly looking for an idea or are just curious. This is what https://t2bot.io/infrastructure/ aims for, as an example.
A lot of people ask either out of curiosity or they want to make a similar deployment. Would be good to have the approximate machine specs in the FAQ somewhere, ideally with a disclaimer that matrix.org is a massive homeserver and you probably don't need half a planet of RAM.