Hacker News Reader

Comment by kondu | Hacker News Reader

I'm sure they considered hosting matrix or zulip, I wonder what the estimated monthly cost of that would be for a community of their size

View on HN

Yeah but what scaling/deployment layer should they deploy it on? I vote for Azure Service Fabric.

at 250.000 users they're going to hit specific limits very quickly, and frustratingly, proper sysadmin skills are (I think) nearly completely eroded from our industry.

This leaves us with expensive offerings despite a pretty static load (a-la; cloud).

Back in the day, burgeoning sysadmins would have cut their teeth on projects like this, but sadly they'd need someone quite senior at this point to avoid major pitfalls.

I'm not even sure myself how I would prevent the abuse of uploaded images; both in terms of rate limiting new accounts and the potential harmful material that might be shared. -- And I am one of the sysadmin types who cut their teeth on problems like these.

For Element, I suspect they'd be best off using Element Server Suite (https://element.io/server-suite) which are the official helm charts for Element, Synapse and the various component parts. To scale elastically they'd need the Pro version, but we could provide them with a discounted license of some kind (but not free, given Element isn't profitable yet and we need the $ to actually work on Matrix...)

If anyone reading this wants to talk, hit me up at matthew at element.io (or @matthew:matrix.org on Matrix)

They haven't considered that, because until this week they didn't need to. Some Linux Foundation projects use Zulip, and the team behind the project seem willing to host for free.

I think consideration may have been limited by the fact that (AFAIK) Slack only provided a week's notice of this change, which has left the Kubernetes volunteers trying to act quickly to avoid losing data which isn't easily archived (private channels and DMs)

Reply on news.ycombinator.com