I didn’t see a single irate email, tweet, or comment on IRC during the entire outage; people politely asked if there was an outage and if it was related to the EC2 outage, and aside from that the only communications I had were very positive, mostly thanking me for my efforts and status reports.
I wonder if this is due to the technical userbase of Tarsnap (which includes myself) that understands that the service is hosted on AWS and that Colin can’t really do anything about it, or whether it’s due to the nature of the service where only people that were trying to do backups at that particular time would have noticed or cared.
Good point. I don’t use tarsnap, but I wonder if Colin actually had a good communications / status update flow going and if that impacted the way users perceived the outage.
The status updates during the outage were close to exemplary – at least for techie guys/girls: http://marc.info/?l=tarsnap-announce&w=2 and http://twitter.com/#!/Tarsnap/. And the story from the other side of the fence: http://aws.amazon.com/message/67457/.