Linden Lab having a bad day
It's not been a fantastic week for the Second Life service (actually more like ten days). Plagued with network problems, the grid has been experiencing everything from strange behavior to partial blackouts of large numbers of servers at once.
Today isn't any better, and actually is looking a whole lot worse.
The graph up there shows concurrency for the last 24 hours. Normally, what you'd expect to see is a relatively smooth sine-wave with a sharp bump at the top as the European users go through peak time. Today starts just before the minimum dip towards the left there.
The 'potholes' you see represent partial service outages - that's assuming the data is being reported correctly - and it seems to be correct. Just after peak, for example, there's a sudden abnormal drop of ten thousand users, consistent with a networking failure between colocation facilities. The sudden jumps upwards represent surges as systems recover and large numbers of people log back in.
Linden Lab's being quiet today, considering, since their notice of a networking issue this-morning. Okay, so it might be considered a bad move to diss your network provider(s), especially when you're expecting them to hop to things and fix stuff for you - but whatever the problem is - whether it's internal to Linden Lab operations, third-party hardware in colocation facilities or third-party networks, I'm sure we'd all like to know what's up.
[Update: 6:10PM SLT - Linden Lab acknowledges and, surprise! Apparently the drama is down to database crashes]
