hello,
downtime during the last ~2 hours (13-15 UTC) was caused by a power strip in the colo blowing. switch/router and some machines were down, but the database is unaffected. the site is now mostly up but may be unreliable / slow while we finish working on it.
we're currently discussing options WRT better management & redundancy to prevent this happening again in the future...
thanks, kate.
(i should point out that this problem was entirely our fault, not the colo's, and they were very helpful in bringing the site back online.
please follow up technical queries to wikitech-l rather than wikipedia-l).