Skip to main content

Probably #Prosody has some problems related to Message Archive Management (including MUCs), my #XMPP server is facing some issues and downtime...

pimux.de/blog/pimux-de-is-faci…

But I am confident I can improve or even fix this 🙂 My monitoring already alerted me a few times but a friendly user told me about the issues, too. That shows me this is a serious problem and I have to do something!

Screenshot of Uptime Kuma dashboard showing that the XMPP server pimux.de has had some outages.

pimux.de is facing some connectivity issues | pimux.de

Grav is an easy to use, yet powerful, open source flat-file CMS
pimux.de
#xmpp #prosody
in reply to Finn

Prosody IM
Prosody IM
• •
In case you're not aware, we have a helpful community chat ( links at prosody.im/discuss ). I'm certain if you need help we can figure it out 🙂

Community and discussion channels – Prosody IM

Various discussion avenues about Prosody
prosody.im
in reply to Prosody IM

Finn
Finn
• •
@prosodyim Thank you, I will check tomorrow, now I have to sleep 😀
@Prosody IM
in reply to Prosody IM

Finn
Finn
• •

@prosodyim I quickly got help, thank you!

There was one user with 131 million records in the prosodyarchive table, perhaps a spam account or a accidental loop as someone/you (?) has mentioned.

Deleting that account and after that doing a VACUUM took about 40 minutes and freed up about 158 GB of space.

A screenshot of a Grafana dashboard showing the CPU usage while deleting a misbehaving XMPP account on pimux.de and later doing a PostgreSQL VACUUM on the prosodyarchive table. It show a really high IO usage over a period of about 40 minutes.
A screenshot of a Grafana dashboard shows that about 158 GB or 16% disk space was freed after deleting one account.
@Prosody IM
⇧