By Harry McCracken | Wednesday, January 27, 2010 at 11:55 pm
For the past few Apple product launches–and other events such as Google’s recent Nexus One phone rollout–I’ve been providing real-time coverage using a free service called CoverItLive. Folks often tell me that Technologizer’s live coverage is swifter than that of larger sites, and I’m quick to give CoverItLive full credit–it’s proven a really good way to post live updates (and photos) with as little delay as possible.
I’ve thought of the service as a secret weapon–but today, it wasn’t a secret at all. The Yerba Buena Center, where Apple held its event, was dense with journalists. A meaningful percentage were liveblogging the news. And a meaningful percentage of those were using CoverItLive, including some great big sites.
CoverItLive couldn’t keep up. A few minutes after Steve Jobs took the stage, I found that my updates were appearing on the site only after a lengthy, atypical delay. Then I couldn’t get into CoverItLive at all. It wasn’t clear at first what was going on–Verizon coverage was unusually iffy–but it eventually became clear that CoverItLive had croaked under the weight of multiple livebloggers and untold numbers of viewers. (We had thousands at Technologizer alone)
Part way into the morning’s proceedings, CoverItLive snapped back to life, and it worked well for the remainder of the event. My thanks to everyone who hung around until we were back in action, and I apologize to everyone who stopped by and found the liveblogging had stalled, or who weren’t able to get into CoverItLive at all. The service has proven remarkably robust in the past, but I’ll have a backup plan in place next time around.
Oh, and here’s a note I received this afternoon from Keith McSpurren, president of CoverItLive:
Needless to say, we failed you and our users today. I take the early support you threw behind CiL very seriously and it’s a very personal issue for me when I let someone like you down.
We did everything we could to prepare for this and had monitoring set up at amazon and rackspace. In the end, we screwed it up on a few servers as we started to double the capacity versus the last one of these.
Thanks for throwing one of the few positive tweets our way.
McSpurren also told me that the company’s working on strategies to prevent widespread problems in the future, such as walling off each user so that overload in one event doesn’t overflow into others. There’s so much that’s good about CoverItLive that I hope the service is indeed able to make today’s meltdown a one-time affair…