The old Daily Kos went off the air at 9 a.m. PT Saturday morning, thought the import would be done by 2 p.m. PT, and that we'd be live by 5 p.m. PT. Ha ha ha ha! These things never go according to plan, particularly since a certain Mr. Murphy never passes up an opportunity to join the party. Heck, importing just the mojo from DK3 took almost two hours. So you can imagine what it must've been like to import nearly one million diaries, 30 million comments, 300,000 users, and countless recommends (plus hotlists, subscriptions, etc, etc, etc).
Once imported, nothing worked. So for several more hours, the tech team scrambled to reconnect all the loose piping. The internets are a series of tubes, you know? Then servers were misconfigured at the data center, yadda yadda, and before we knew it, we didn't go live until 1:30 a.m. PT.
Yay! We were in business! Except that once the West Coast woke up in the morning, the crush of curious onlookers overwhelmed a brand-spanking new engine untested by real-world traffic. Under the load, it all but collapsed. Some users griped about the poor "first impression", others were more understanding of the nature of this thing. All the while, the tech team furiously worked to get the site functional, and by early afternoon PT, they had succeeded.
But they're not out of the woods, not by a long shot. Today, we have the next big test -- much bigger weekday traffic.
Good news, the site instability culprit has been identified (has to do with the site's caching). Bad news, it'll require a fairly extensive rewrite of that module, so it's not a quick and easy fix, and the longer the tech team is working on core stability, the longer it'll take them to tackle some of the smaller but annoying "papercut" types of bugs. But so it goes.
That said, you guys were sure busy on Sunday. As of 9 p.m. PT, when I wrote this piece, there had been 177 groups created, 289 diaries written, and 12,933 comments made. Remember, that was for about 19 hours. Not a full day. To compare, 168 diaries and 14,827 comments were made all of last Wednesday -- a day with 60,000 more visits (again, as of the writing of this post).
So our short-term plan is to survive Monday's higher traffic. The tech team will be doing less bug squashing, and more "staring at server loads ready to pounce when things slow down". And they will slow down. Don't be surprised at the occasional outage and error message. After those problems are squashed for good, then the programmers can turn back to fixing bugs and upgrading the user interface in those areas we see are creating confusion.
All I ask is that you bear with us as the tech team fixes stability issues and the myriad of annoying bugs that still remain. Don't worry so much about bug reports -- at this point, I suspect we know them all, and it's a big enough list. But if you have user interface suggestions, tag your diary "DK4", because I follow and read every one of those diaries. Or find some of those diaries and add your suggestions in the comments, because I don't just read the diaries, but all the comments as well. As the beta testers will attest, we've made a ton of changes based on user feedback, and we're nowhere near done making changes.
Finally, a reminder:
Joining a group means [running it]. That's by invitation.
Following the group merely funnels its diaries into your stream. That's optional, if you never want to miss a group's diaries.
So there's the crowd that plans the party, the crowd that signs up for the invite list, so they know when and where the party is happening, and then there are those who merely show up when they come across the party.
So to be clear, you only join a group if you're going to help run it. For 99.99% of you, that's not anything you need to worry about.
Update: The mobile site will be fixed by the end of the week. It'll be like the last one -- limited functionality. We'll roll out a fully-featured mobile site later this year.