The outage appears to have begun about 2 p.m. PDT Wednesday. "We're aware of a problem with Gmail access affecting a small number of users," said a Google moderator named Mark in a Google Groups post on Wednesday. "We expect to provide an estimated problem resolution time at 3 pm [PDT]. Please note that this is an estimate and is subject to change."
The projected resolution time was subsequently adjusted to 6 p.m. PDT on Wednesday.
On Thursday just before 1 p.m. PDT, Mark posted an update. "The issue affecting a small number of Gmail users is being addressed, and many users have already had their service restored," he wrote. "All affected users should have their service back online today, if not already restored. We know how important Gmail is to our users, so we take issues like this very seriously, and we apologize for the inconvenience."
The final update, indicating that the problem had been resolved, came at 6:12 p.m. PDT on Thursday.
In the Google Apps Discussion Group, on Google Groups, someone posting under the name Bill W., who apparently plays a role in IT administration for Alachua Public County Schools in Gainesville, Fla., expressed frustration Thursday morning with his inability to escalate the issue for faster resolution. "Since yesterday around at least 4 p.m. my CEO cannot access his mail," his post says. "He gets a 502 temporary error. Support keeps telling me it is affecting a small number of users. This is not a temporary problem if it lasts this long. It is frustrating to not be able to expedite these issues."
This, of course, is the downside of any outsourced service, as vendors of on-premises software gladly point out. A Microsoft representative, for example, contacted InformationWeek's Eric Zeman to make sure the Gmail outage didn't go unnoticed.
In July, when Amazon Web Services' Simple Storage Service (S3) suffered a service failure for about eight hours, some S3 customers expressed similar worries about the suitability of S3 for businesses. Others dismissed S3's various outages as "few and far between, short, and handled properly."
Google's Gmail service-level agreement (SLA) for paying Google Apps customers says the company "shall use all reasonable commercial efforts to ensure that the Gmail Web interface is operating and available to customers 99.9% of the time in any calendar month."
October has 31 days, or 744 hours. To achieve 99.9% uptime during October, Gmail would have to be available for at least 743.256 hours, or down for no more than 0.744 hours (44.64 minutes). Because the outage appears to have lasted more than 24 hours for some, affected paying Gmail customers appear to be owed service credits, as per the terms of the Gmail SLA.
Google did not immediately reply to a request for comment.