| 9 Dec 2024 |
 richvdh | ok, so to confirm: that would correspond to 21:54 UTC? | 22:15:10 |
 temporarynep | correct | 22:15:27 |
 TravisR | I'd rather not, tbh. It's already beyond my comfort level to sign myself up for this many pings 😅 | 22:19:19 |
| richvdh | Having poked at the logs a bit, I think it probably is the same problem as https://github.com/element-hq/synapse/issues/17035, or at least something very much like it. Again, I see the request in matrix.org's haproxy logs, but Synapse drops the connection before the request completes. | 22:41:52 |
| temporarynep | thank you for checking <3 | 22:45:10 |
| temporarynep | since the query from the comments didn't seem to fix even temporarily, i guess now i have to figure out what to do, | 22:49:13 |
| temporarynep | * since the query from the comments didn't seem to fix even temporarily, i guess now i have to figure out what to do | 22:49:17 |
| richvdh | it's probably worth a go restarting synapse after dropping the messages, if you haven't already tried that | 22:50:44 |
| richvdh | (synapse will keep retrying the same transaction for a while) | 22:50:55 |
| temporarynep | i ran it with synapse fully off actually, since i wasn't sure about how it handled safety in that regard | 22:52:29 |
| temporarynep | 1.120.2 fyi | 22:52:39 |
| temporarynep | * i ran the query with synapse fully off actually, since i wasn't sure about how it handled safety in that regard | 22:53:40 |
| richvdh | hrm :/ | 22:55:30 |
| temporarynep | oh neat, removing the size check (i.e., running delete from device_federation_outbox where destination = 'matrix.org';) seemed to work (somehow)! | 23:48:21 |
| richvdh | right, so yes. I looked at your traffic | 23:57:52 |
| richvdh | the problem was not 500000-byte messages; it was, rather, a large number of 400000-byte messages | 23:58:38 |
| richvdh | that's going to mess up a lot of outbound encrypted messages :/ | 23:59:16 |
| 10 Dec 2024 |
| temporarynep | it's not too bad since my server is mostly just for me, my bf and a few friends | 00:00:14 |
| richvdh | you must have been sending messages to some very big encrypted rooms. | 00:01:42 |
| richvdh | anyway, glad you've fixed it. I believe the backend team are looking into this. | 00:02:27 |
| temporarynep | lol yep, there actually seems to be a vodozemac issue that causes encryption to take far longer than it did with libolm in big rooms, up to multiple minutes (reproducible with all vodozemac-based clients i tested, including element android + web and fractal) but that's neither here nor there anyway, i should look into reporting it on github if no one has done it yet | 00:03:38 |
| temporarynep | and thank you | 00:03:47 |
| temporarynep | * lol yep, there actually seems to be a vodozemac issue that causes encryption to take far longer than it did with libolm in big rooms, up to multiple minutes (reproducible with all vodozemac-based clients i tested, including element android + desktop and fractal) but that's neither here nor there anyway, i should look into reporting it on github if no one has done it yet | 00:07:23 |
| temporarynep left the room. | 00:07:33 |
|  Yorusaka Miyabi changed their display name from Yorusaka Miyabi | Emotional Bugged | No DM to Yorusaka Miyabi. | 06:21:46 |
 bkil | @nex (she/it) I invited you because we were talking about v1. And as I myself reject DM by random people, I usually avoid sending out random DMs for the same reason. 🤷 | 12:38:14 |
 nex (she/it) | Ah okay | 12:50:06 |
| nex (she/it) | just gimme a ping somewhere next time :^) | 12:50:24 |
 ninchuka | yeah simply asking here if she would like to be invited to a room where thats on topic would not require this | 12:52:33 |
 Emma [it/its] | i too am now having issues with matrix.org over federation...? | 13:33:34 |
