UPDATE (2011-06-02): The switchover to our content delivery network is complete. Now, all images are loading from a worldwide network of servers for faster loading. The site is now measurably faster for both logged-in and anonymous users and better able to handle the large spikes in traffic we get after major releases. Thanks for bearing with us while we made the transition and enjoy the new speed! - John
=== Previous Updates ===
(2011-05-27) I've begun the switchover to our new content delivery network for much of the static images on the site. App logos (16px, 75px, 128px), app screenshots (800px and 200px), homepage graphics and sitewide header/sidebar/footer images have all been switched over to the CDN. This required scripting to update approximately 1,800 pages and page revisions on the site. I've reset our caching system and its rebuilding now. Old news posts for about a dozen apps may be missing app logos due to some collisions in naming.
If you notice any issues with any pages in the app directory missing an image for an app logo or screenshot or having the wrong image, please post a comment about it here and we'll get it fixed. Things should be settled by late morning Friday NY time.
UPDATE (2011-05-30): Nearly all the rest of the images on the site have been switched over to the new content delivery network. The cache for the ~200,000 pages in the site is being rebuilt but as it is a national holiday in the US, most folks shouldn't notice too much of a slowdown. The rebuild should complete later tonight when we'll be doing a bit of server maintenance and taking the server offline for about 10 minutes.
Visit the Community page
Join our forums
Subscribe to our email newsletter
Subscribe with RSS
Follow us on Facebook
Follow us on LinkedIn
Follow us on Twitter
Follow us on Mastodon
Follow us on BlueSky
The category icons, site buttons and other site elements from the footer, sidebar, stories and app directory have now been switched over to the content delivery network. As it's a slow day here in the US (national holiday), traffic is a lot lighter, so the additional load to rebuild the cache for all the pages won't be as noticeable.
Sometimes, the impossible can become possible, if you're awesome!
The final images that fell through the cracks have been moved and the site is performing very well. The cache was rebuilt again yesterday afterwards. We've disabled pipelining as it's less useful now -- no images to load and one set of CSS and JS to cache in the browser -- and that's upped our parallel user numbers, too.
I was also able to update the maintenance page while running our updated backup timing tests to specify when scheduled maintenance is occurring (1am to 1:10am NY time) so users are less puzzled by seeing it mid-day in Australia and other places on that side of the world. For the unfamiliar, our site does a database backup nightly at 1am NY time and it takes about 15 minutes. Well, it used to, now it's about 10 with moving the images to the CDN. During that time, most guests should still be able to see the site. Many logged-in users will get the maintenance page (due to the database being overloaded) but at least now they'll know why.
Sometimes, the impossible can become possible, if you're awesome!