Love the lack of propper media cache tooling on mastodon.

No way to clear whatever just pushed over 4GB of media across the fediverse

are there any mastodon forks that don't cache media

(and i really mean mastodon. i am already very familiar with *oma, misskey, etc)

#fediadmin #mastoadmin #mastoadmins #fediadmins #fedimins #mastodon ​​​​​​​​​​​​​​

Mastodon admins currently getting slammed by Earth's least creative spammer really shows how awesome #mastoadmins are if you haven't seen this spam yet.

But bruh we need more spam fighting tools. This spammer is literally sending identical DMs to probably thousands of users and we don't have a stock tool that allows us to auto block/flag it.

I've been experimenting with enabling translation on my instance and since I've seen zero documentation about how to configure LibreTranslate with a purchased api key, I'm going to share my findings here.

Pretty much every guide out there for enabling translation on a Mastodon instance using LibreTranslate assumes that you are going to be setting up the LibreTranslate server yourself. That's admirable, but translation is pretty resource-intensive. When I tested this on my instance, it was taking about 2 seconds to translate a post, and even longer for longer posts. Plus there were some languages that weren't working at all. That was probably on me for not setting up the language packs correctly. Maybe I'll give it another try, but somewhere in the middle of wrestling with the translation server, I realized that I'm here to run a Mastodon instance, not a LibreTranslate instance, so I went and purchased an api key from LibreTranslate for their lower tier of $29 a month (good for 80 translations per minute). Based on the average donations we get each month, that seemed a reasonable cost that we could afford. The trouble was, configuring the instance to use the api key didn't seem to be working at all.

Now if you go to https://libretranslate.com it lets you play around with the api and shows what the post request and the subsequent response looks like. In the post request and in the api documentation, the endpoint for the translation service is:

https://libretranslate.com/translate

but if you set LIBRE_TRANSLATE_ENDPOINT to that value in your .env.production file, it won't work. After a bunch of googling and experimentation, I finally went and looked at the pull request for this feature on GitHub. And that's where I saw that in the code for the post request it takes the configured LIBRE_TRANSLATE_ENDPOINT value and then adds the "/translate" at the end of it.

Even though I was sure I had tried this before, I set LIBRE_TRANSLATE_ENDPOINT to:

https://libretranslate.com

Without the "/translate" on the end. I restarted the web service and cleared the cache and it started working perfectly.

It'd be nice if any of this was actually documented somewhere.

Any #mastoadmins have issues upgrading from 4.3.2 to 4.3.3? I seem to be missing a bunch of gems. Followed the usual steps.

Hey #mastoAdmins

Is there an easy way to force #mastodon to to be more aggressive in backfilling comments? One of my least favourite aspects of running indie is having to open posts in the browser to see all the comments.

Like, if someone I don't follow posts, I can see a few comments. When I open in browser, I can see many more - including some from my mutuals (that did not display on my instance).

Has anyone done an install of Mastodon built on top of AAPanel for Nginx, Redis and Postgresql management? Thinking of migrating servers and looking to make some of the process easier.

Any #mastoAdmins encounter 403 errors with #libretranslate, running both on same server (not using an API key)

LT web works, and I can make API requests from curl, but #mastodon is throwing a 403.

*insert that Bernie Sanders meme here where 'Once again' he asks us to support our Mastodon instance admins*

https://moth.social/@mammoth/113646767601480718

I'm seeing a lot of HTTP response code 429 (Too Many Requests) on the media_proxy endpoint, which I guess is a known issue but seems silly none-the-less.

Finally updated my instance for the first time in...a while. Does anyone have a good #prometheus exporter config for #OpenTelemetry? I was relying on #statsd before.

what's wrong with mastodon.sdf.org? And who is TomCox? I suspect TomCox is flooding every instance (at least gts instances) with their requests

#mastoadmin #mastoadmins #fedimin
https://fedi.stfn.pl/@stfn/statuses/01JE48BZ6MAH5QB0MEWHBSWWDN

#MastoAdmins's can easily help the #Tor network by installing the Tor daemon on the linux server and configuring their web server to be a #WebTunnel bridge

disobey.net does it (https://metrics.torproject.org/rs.html#details/7E12D6CACBA1051FD23A48482E091257CD4701BC), you can too!

https://www.bleepingcomputer.com/news/security/tor-needs-200-new-webtunnel-bridges-to-fight-censorship/

"Hiding in plain sight: Introducing WebTunnel"

https://blog.torproject.org/introducing-webtunnel-evading-censorship-by-hiding-in-plain-sight/

Hello #MastoAdmins and #MastoDevs

Let's say my database was thoroughly hosed.

If I was to export my profile, nuke my current instance and stand it back up on the same domain and import my profile again, would that work?

Or is the instance backing domain changing likely to cause weird side-effects talking to the other servers?

That's a new one. Any thoughts?

(Mp4 upload)

Attention Beige Party-goers!

Ok! Streaming service appears to be working now! Long story short, I think I misinterpreted some of the upgrade steps for 4.3.0. One of the major changes is that the streaming service got moved over to a separate docker image. Well I updated the yaml file to point to the new image, but I missed a step because the build for the streaming service also has to point to a separate streaming docker file.

Ok, long story even shorter: pretty sure the streaming service was broken since the last update, but due to the nature of docker containers it was continually restarting, so it would work in fits and starts and then die again.

I did hear some reports that feeds seemed to be "stuck" and then would update after several minutes, so that would seem to fit with that theory. Weird thing is that every time I checked myself it seemed to be working, but I didn't go into docker and actually check the status of the streaming service. I have done so now and everything seems to be up and healthy, so I think we are good.

Anyways, welcome to version 4.3.1, and thanks for bearing with me while I figure all this out!

Beige-bless

PS: If any #MastoAdmins, are having this issue, this is where I found the solution:

https://github.com/mastodon/mastodon/issues/28828

vault37.xyz has been updated to 4.3.1!

Question for #MastoAdmins. I just did the upgrade to 4.3.0 and everything seems to be working except for notifications. I ran the pre and post deployment database migrations, and the old notifications got pulled over, but no new notifications are showing up. Anyone else have this problem?

Editing this because I was able to resolve this issue by rebooting the web server. I'll leave this post up in case anyone else has this problem.

Hurrah! I've greenified the PKU Talk mastodon server - massive thanks to @TangerineUI - the code was so nicely organised, it was a breeze to CTRL-F relevant colours and replace.

Now thinking about making a tool to help generalise Tangerine UI into something users can provide their own colours...

You can get Tangerine UI here for your Mastodon server: https://github.com/nileane/TangerineUI-for-Mastodon

#Mastodon update to version 4.3.0 successfully completed! This update requires a bit more attention than we #mastoadmins are used to, but if you follow the excellent #instructions, everything should go smoothly.

https://github.com/mastodon/mastodon/releases/tag/v4.3.0