It looks like there's a wave of spammers going "hey, let's collaborate" and "ooh, we want to verify your account". I've caught up on suspending those idiots. Thx for the reports!
| Over on Metafilter | https://www.metafilter.com/user/326460 |
Pronoiac, with admin hat
- 218 Followers
- 1 Following
- 266 Posts
For timely matters, cc @pronoiac.
Clearing out another wave of those bot spam accounts. It looks like hf.space is related to huggingface and AI experiments, so I'm blocking that domain wholesale.
New emojis, loaded from troet.cafe:
*
- rainbow sheep
*
- blobcat, holding a teddy bear
*
- rainbow sheep*
- blobcat, holding a teddy bearHeads up, I ran across a couple of accounts with apparently broken images, starting a week or two ago. Upon investigating, it looks like the issue is somewhere with our hosting; I've reached out to hosting.
The accounts I saw with issues were on freeradical.zone. Let me know if you see others.
Adding some emoji, let's see if they work:
* 
- Discord, via tech.lgbt
* 
- Anthropic / Claude
* 
- ChatGPT
* 
- Google Gemini
* 
- Microsoft Copilot
Heads up, I just suspended ravergram dot club. There's a serial troll, Jimmy Truth, that landed there, and looking for the administrator, I couldn't find an active account, so, apparently unmoderated. It's also on the Oliphant blocklist; I'm not the first one to notice.
Happy holidays, everyone!
I don't have a feature list prepared yet; I'll work on it.
I doubt keyword search will happen; we're already on a version that supports it, but our hosting hasn't enabled it, likely due to the resources needed.
There's work in progress on bringing Hometown, the fork of Mastodon we use here at mefi.social, up to date. There's a chance we'll jump from v4.2.17 to v4.5.x.
We have uptime checking, which didn't fire an alert. I'll need to check that and tune it.
We have a status page - https://status.mefi.social/ - but it was borked, as the machine that was on, cloudlet, had its own issues. Amazon / AWS / Route53 handles the DNS resolution for our instance; other machines resolve from there, and Amazon had its own issues yesterday. I investigated DNS settings from cloudlet, everything looked *terrible,* and it turned out the DNS was borked on cloudlet. Cloudlet's feeling much better now.
I started an incident report in our Github repo - https://github.com/pronoiac/mefi.social/blob/main/incidents/mefi.social-2025-10-20.md - though I haven't closed and updated it yet.
Setting up a wiki *might* be better, and easier to remember and use.
Our hosting dashboard can show a bit about the tasks backlog, and I can look at Sidekiq now, and both seem fairly caught up.
So we're good for now, with some new items on the to do list.
