We're aware of a problem with Gmail affecting a small subset of users. The affected users are unable to access Gmail. We will provide an update by February 24, 2009 6:30 AM PST detailing when we expect to resolve the problem. Please note that this resolution time is an estimate and may change.
From: http://groups.google.com/group/Gmail-Help-Announcements-and-...
That's a pretty large small subset. Everyone I talked to this morning, in western europe, eastern europe, and australia (and I in the US), was unable to access gmail.
The front page does not show recent submissions, but 'new' does. The front page uses an algorithm based on the 'acceleration' of postings, their age and total points to figure out what to display.
Just recently installed (Downloaded? Activated?) offline mode. So glad I did. I can't wait for Google Calendar to start this, then I'll have outlook effectively replaced.
Reading this drives everyone to immediately check his account further increasing the load on the servers. Maybe we should sit back, relax and remember the times when email did not yet exist.
It's not a load problem, most likely. I think Google is well equipped to deal with load spikes, and in any case, we'd see a more gradual failure. This was pretty sudden.
Also quite funny to see a thread on this comments page complaining about a single duplicate HN submission, when twitter has thousands of "Gmail is down" duplicate tweets.
The compression ratio on twitters db must just be astounding.
A friend of mine is reporting the same problem down here in Honolulu (on Clearwire wireless) but I'm on RoadRunner (also in Honolulu) and all of Google's services work just fine for me.
So yeah... is this really necessary? Everyone and their brother uses Google/Gmail, the only thing you're doing here is letting everyone know you're were the first on the scene. But, you probably weren't, and the first guy on the scene didn't care.
I was referring to the "try again in 30 seconds". When your service is failing, maybe it's not the best time to invite users to pummel it with requests. I think a generic "try again later" would work better.
