I hadn't planned on making my blog a sounding board for all products that have had significant downtime, but this one has certainly hit close to home.

Starting yesterday evening, around 8 p.m., I have been completely unable to add new posts to the site, making it appear that I am asleep at the wheel. The culprit? An issue with Google's Blogger service, which has blocked the ability to post via FTP.

This is not the first time Google's Blogger has had a outage of significant length here, and also, not the first time they have completely ignored a throe of user complaints and support requests on their site.

At a time when Wordpress and other platforms are gaining significant momentum, and can tout "5 minute" upgrades, the temptation to move, assuming the site structure and comments are retained, is extremely high.

I'd have thought Google's acquisition of Blogger via Pyra Labs would have provided the team with significant experience in growing a scalable, trouble-free infrastructure, but from conversations I've had with people close to the team, it seems that the most infrastructure-focused employees at Blogger had stars in their eyes around Google's other products, and Blogger has suffered from neglect.

Something is Broken indeed.

Update: Finally acknowledged on Blogger Status and FTP is starting to flow. But this is ... bad.

See Also:

Google Blogger: Something is Broken!
http://tinyurl.com/5sejhy

Thomas Hawk Cites Blogger Outage on FriendFeed
http://tinyurl.com/3p96y8

From August 9, 2007:
Google Ignores Users During Major Blogger Outage
http://www.louisgray.com/live/2007/08/google-ignores-users-during-major.html

While some out there are looking to refresh their RSS feed subscriptions by starting from zero, I believe you can still get the benefit of new voices by selectively adding individual feeds, without getting overloaded.

In the vein of the last three months' offerings of new blogs you're likely not reading, (March | April | May), below are five more bloggers who I think are worth taking a chance on, whose writings have caught my eye of late, and who need a bigger stage.

1) Coding Experiments (http://codingexperiments.com/)

Focus: Technology, Development, Coding, Social Media
Recent Highlight:
How Many Features Can Be Cut from a Service without the Users Leaving?
RSS Feed: http://codingexperiments.com/feed

2) LiveCrunch (http://www.livecrunch.com/)

Focus: Social Networking, Plurk, Twitter
Recent Highlight: Things To Do Before WWDC 2008
RSS Feed: http://feeds.feedburner.com/livecrunch/technology

3) Broadcasting Brain (http://broadcasting-brain.com/)

Focus: Social Media, Podcasting, FriendFeed
Recent Highlight: Punching above your weight with social media
RSS Feed: http://feeds.feedburner.com/BroadcastingBrain

4) Neoformix (http://www.neoformix.com/)

Focus: Statistics, Data, Twitter
Recent Highlight: Top Twitter Users StreamGraph
RSS Feed: http://www.neoformix.com/index.xml

5) Michael Fruchter (http://michaelfruchter.com)

Focus: RSS, FriendFeed, Social Media, Toluu
Recent Highlight:
Cleaning up my Google reader with the help of Toluu.
RSS Feed: http://michaelfruchter.com/blog/feed/

As mentioned last month, I'm always looking for more new bloggers and interesting voices to be added to my Google Reader feeds. You can get an early tip as to new bloggers I'm following and sharing by signing up for Toluu, following me on FriendFeed, or following my Google Reader shared links blog. If you think there are more I should check out, please leave them in the comments.

It seems that outages are the new black.

After a weekend filled with stories on Amazon downtime, a brief Disqus blip, continued Twitter troubles, and many sites straining to take on increased crowds swelling to catch the latest from WWDC, I was surprised to see my blog statistics engine, SiteMeter, get in on the act. Since 11 this morning Pacific time, data has been almost completely stalled, not logging visits, and the company's blog doesn't give any reason for the slowness.

I'd like to blame Scoble, or blame Steve Jobs, but I don't think they're the cause.


SiteMeter is one of the most widely used statistics trackers in the blogosphere. And while I could put up with occasional outages from a free product (See: Mark Evans: The Wonderful World of Web 2.0 Whining), I'm one of those who wanted to support the site's developers, paying $89 a year to gain a premium version of the service last year, which gave me expanded access to a wider array of reports.

I'd like to say I don't check with SiteMeter throughout the day out of curiosity, but I'd be lying to you for sure. I love stats. I even made a dashboard widget for Mac OS X that shows me the day's activity, letting me just drag my mouse to the bottom right corner to get caught up. Except, today, I was surprised to see I was extremely unpopular. Not only was the total visit count much lower than I had anticipated, but it said absolutely nobody had checked in in the last hour. And since this morning, I've seen no updates at all.

SiteMeter's seen issues like this in the past. They operate not from one mega-database, but instead, each of its individual servers runs on its own database. When one has a hiccup, only those users on that single server show issues. I expect that's likely what's going on here, and just maybe, with luck, the total statistics will catch up overnight.

Now, we'll see just how much my going dark for about 36 hours over the weekend will have hurt me. With the company's blog not giving any hints as to what's happening, hope is all I have. Maybe it's time to check in with my FTP server and download my logs.

Today, many of the popular sites aimed to deliver minute-by-minute updates to Apple CEO Steve Jobs' keynote, as well as some social networking sites extremely popular among the technology elite, slowed to a crawl under the rush of traffic from Mac and iPhone fans hoping to get a glimpse of Cupertino's latest products. Sites as diverse as TechCrunch to Twitter buckled under the pressure, while others, like MacRumors Live, Engadget and FriendFeed, maintained stability, gaining praise.

I am happy to report that louisgray.com enjoyed 100% uptime during this rush. Here's how we pulled off the enviable feat, all without removing services, reducing features, or requiring the offloading of some traffic to partners:

1. I Posted Absolutely Nothing At All

After much advance study, I realized that one of the major issues behind some of these sites who ran into trouble was in their offering of interesting content. Whether through rumors of new products in advance of the conference, live feeds during the conference, or reaction to the conference's announcements, each of the sites had attracted a population of users disproportionate to the norm, resulting in traffic spikes well above average, invariably causing slower access times or even downtime.

To avoid such a fate, not only did I not promise anything, but I didn't post anything, not just the morning of the keynote, but in the preceding 24 hours, essentially throwing potential visitors off the scent. I believe that this strategy, delivering an overall reduction, both day over day and week over week, in terms of total visitor traffic and page views, left the site with considerable headroom, and reduced chance of service interruptions.

2. I Made No Modifications to the Infrastructure

For several years now, louisgray.com has been hosted via FTP on Register.com hosted servers, powered by Google's Blogger engine. After considering many options available in advance of the WWDC keynote, it was determined the best course of action was again, nothing. Given the site's near 100% uptime over the last few years, despite significant year over year growth, the prevailing bias was to hold off on any significant software or hardware purchases which could cause complexity.

3. I Made No Advance Promises to Uptime

Murphy's Law dictates that anything that can go wrong will go wrong. Many were surprised as to Twitter's advanced promise of significant uptime during the keynote, after such a recent spotty track record. By promising 100% uptime for louisgray.com, I knew that I would, in turn, be placing myself in the line of fire for overzealous hackers and an overcaffienated faction from the Mac army, ready to take my site down like so many others.

Conclusion

I think there's no other option except to congratulate myself for delivery of 100% uptime during a time of considerable stress for tech media giants. Where they zigged like moths to the flame, I zagged away from the noise, bravely hiding in the corner, cowering in fear. You can count on louisgray.com to deliver the 100% uptime during such mega-tech events, both now and in the future thanks to our unique strategy of reverse traffic optimization. I hope we can count on your support.