archwiki: Fight load spikes caused by crawlers
-
tf-stage1: Upgrade the Wiki box to CPX52➡️ Give the server more resources -
archwiki: Write Mediawiki specific anubis rules➡️ Enable Anubis selectively for expensive endpoints such as "Special:RecentChanges" etc. -
anubis: Create config dropin structure➡️ Create a config structure so that we can more easily configure anubis on a per-service basis. -
anubis: Drop paid-only configuration➡️ Drop config which requires thoth service (paid)
Undone again:
-
archwiki: Enable Extension:CrawlerProtection➡️ make the expensive pages only available for logged in users Extension:CrawlerProtection
Cherry-picked to master:
archwiki: Fix the error_log in proxy location
See https://www.mediawiki.org/wiki/Manual:Handling_web_crawlers
cc @archlinux/teams/wiki/admins regarding the wiki config changes, as for example page history is only viewable for logged in users right now.