Puella Magi Wiki talk:Maintenance Log
~~~~) at the end of your comment.TemplateStyles & CodeEditor
I've added TemplateStyles and CodeEditor to the wiki, so styling the wiki could be less suffering. CodeEditor should be enabled automatically when editing non-main namespace pages with .css suffix (e.g. Template:Foo/styles.css). Though, the page may need to be changed to "Sanitized CSS" via Special:ChangeContentModel before they're usable in <templatestyles src="..." /> tag. -- 0x99 (talk) 20:23, 15 September 2023 (UTC)
Email notification
If you've confirmed an email, email notification for watchlist and talk page can now be enabled (in Special:Preferences). I am not sure how well this works in practice (spam and stuff), but let's see! -- 0x99 (talk) 18:39, 25 September 2023 (UTC)
Downtime
Looks like there were some issues with the server (on the managed hosting side) and the wiki went down for a few hours. Sorry about this. If it happens again, I'll consider moving us to a VPS again. -- 0x99 (talk) 16:53, 7 October 2023 (UTC)
- I've migrated the server to a VPS, so the wiki should be much more stable now. I'm continuing to fix things, so if you find any stuff broken, please report back. Thank you! -- 0x99 (talk) 21:19, 27 October 2023 (UTC)
Downtime (2024/07/15)
Hetzner VPS node was down for about 30 minutes due to an error on the Hetzner side. This has been resolved. -- 0x99 (talk)
Downtime (2025/04/09)
The server was running out of space and went down for about 9 hours. I've resolved this by upgrading the server to a higher tier (more CPU/more RAM/more disk). -- 0x99 (talk) 14:30, 9 April 2025 (UTC)
Downtime (2026/03/03)
You may have noticed the website has been very slow to access lately. The wiki has been experiencing a possibly targeted bot attack in the past few weeks (and possibly longer). I've attempted a lot of mitigation, but none seem to effectively fend them off for more than a day. I've now migrated the server to use Cloudflare for attack protection. Hopefully this lasts. -- 0x99 (talk) 14:41, 3 March 2026 (UTC)
- Better prepare to monitor the source of those attacks and launch more aggressive Javascript challenges for those countries/useragents. What was being used before Cloudflare? ImperviousCover (talk) 08:54, 4 March 2026 (UTC)
- The Wiki has been relying on Varnish Cache and manual IP/UA blocking to fend off these attacks in the past, which has worked very well. This time, the attack specifically DoS'd some of the heaviest pages while managing to bypass the cache. Cloudflare seems effective so far, as I can see the loads are back to the normal levels today. -- 0x99 (talk) 13:38, 4 March 2026 (UTC)
- Yes, for now it should work. I don't know how "famous" this site is and thus how many requests it has and how distributed they are, but you most probably are going to have to monitor requests from time to time and change some restrictions at some point. I don't think you are going to require more than this though. ImperviousCover (talk) 14:42, 4 March 2026 (UTC)
- The Wiki has been relying on Varnish Cache and manual IP/UA blocking to fend off these attacks in the past, which has worked very well. This time, the attack specifically DoS'd some of the heaviest pages while managing to bypass the cache. Cloudflare seems effective so far, as I can see the loads are back to the normal levels today. -- 0x99 (talk) 13:38, 4 March 2026 (UTC)
Downtime (2026/03/04)
The server went down for 5 minutes today due to applying new network-level tunings. Sorry for the inconvenience. -- 0x99 (talk) 17:09, 4 March 2026 (UTC)
Slowdowns (2026/08/13)
AI crawlers have been slamming the server pretty hard on a few heavy pages that cannot be cached. We've now switched to heavily throttling those pages for crawlers. The wiki should now be performant again. Sorry for the inconvenience. -- 0x99 (talk) 16:40, 13 August 2026 (UTC)
Slowdowns & Cloudflare 525s (2026/08/17)
The wiki is currently experiencing some slowdowns, as well as Error 525 on Cloudflare. I have done some extra tunings to fix the slowdowns (which require a FreeBSD upgrade due to package end-of-support). For error 525, this may take a while to fix. I suspect it may have more to do with the networking side than the server configuration. -- 0x99 (talk) 17:55, 17 August 2026 (UTC)
- Temporarily disabled Cloudflare; and all these weird instability are gone… -- 0x99 (talk) 17:59, 17 August 2026 (UTC)
- The wiki constantly breaks down now ;-; Is it because of Cloudflare being disabled? Because I was not gonna write here but its gotten bad T-T. - TheresaFrog421 (talk) 15:53, 19 August 2026 (UTC)
- Sorry, this was an unrelated issue; the server was running low on disk space for a while, and when the automated job tried to create a backup archive, the temporary files managed to fill the entire server's disk in the way that crashed everything. This should be resolved now with more resources. -- 0x99 (talk) 17:23, 19 August 2026 (UTC)
- p.s., thank you for the ping. I have some monitoring set up, but it's only hourly, and can fail to alert if it's flip-flopping between working/not working. Commenting here helps a lot! -- 0x99 (talk) 17:39, 19 August 2026 (UTC)
- The wiki constantly breaks down now ;-; Is it because of Cloudflare being disabled? Because I was not gonna write here but its gotten bad T-T. - TheresaFrog421 (talk) 15:53, 19 August 2026 (UTC)
Downtime (2026/08/19)
The server was down for an unknown number of hours due to the disk space being full. The server was running low on disk space for a while. When a regular backup created temp files for archival (e.g., DB snapshots, Redis snapshots), it managed to fill in the remaining disk space in the way that crashed the entire server. This should be resolved now that I've provisioned more resources. (We'll also need to scale up server resources overall, as Walpurgisnacht has resulted in quite an upward tick in traffic.) -- 0x99 (talk) 17:33, 19 August 2026 (UTC)