BuzzFeed’s website went offline Monday night (as did other news sites like Huffington Post and Gawker) when the data center housing its servers flooded. Pando Daily’s David Holmes talked to BuzzFeed editor Ben Smith about how the site responded — switching all publishing over to Tumblrs as a stopgap while rebuilding its own site, from scratch.
Just three developers worked throughout most the night to get Buzzfeed.com back up and running [in the cloud on Amazon Web Services]. One of them, Eugene Ventimiglia, kept working even after a tree fell through the roof of his home in North New Jersey.
“It took years to build (Buzzfeed) and they rebuilt it in six hours,” Smith said.
Of course, AWS cloud hosting has had its own failures when weather or power outages affected its server farms in Northern Virginia. So it’s probably smart for news orgs to have layers of backup plans.
There can often be a virtual wall between the editorial side and technology side of a news organization. Newsroom editors may need to start asking more questions about their site’s technology setup. How and where is our website hosted? How is data backed up? How would it be restored, how long would that take and what would it look like as that process was under way? Are there redundancies in case one part fails?
A newsroom should also have a plan for how to adapt when the technology does fail. Last year I wrote a step-by-step guide to planning for how to get news to your audience during a website outage.
Some of the most-important steps to take before a potential outage include these:
- Decide what services you will use. Tumblr seems to be a primary fallback now for many news orgs. The Huffington Post used Blogsmith when its site went down during this week’s storm. Be prepared to use Twitter and Facebook to share information and get the word out.
- Assign individual responsibility and authority. Who decides to flip the switch to the backup system, and how? Who is responsible for letting everyone know?
- Set up the backup websites and accounts in advance. You can keep them private so users don’t get confused by stumbling across them. But best to have these ready to go when you need them. Or, if you’re already using your Tumblr for something else, you have a built-in audience ready to find you there.
- Include a prominent notice to readers. Let them know the main site is down and this is a temporary fix.
- Link to the last available cached version of your site. If your own servers or Google’s have preserved copies of what was published previously, help readers find it.
- Train the staff. If your backup plan is Tumblr or some other alternative publishing platform, make sure your staff knows how to use it before they have to start publishing on it.
It’s an easy thing to put off, but taking some time now to make your plan will leave you and your readers much more prepared when the next storm approaches.