Facebook’s Staggered Maintenance Procedure

Facebook has one of the world’s largest server farms and for good reason – with all that traffic no amount of servers can be considered too much. While Facebook’s uptime is a lot better than many most of the other “Web 2.0” services, we’ve been seeing a lot of maintenance-related downtime recently (see possible reasons below).

You’ve got to hand it to the Facebook team though, they have scaling and uptime perfected down to an art. For instance, when servers are due for updates, the maintenance is performed in a staggered manner, updating one set of servers at a time as attested to by the unavailability of certain Facebook accounts while others can still be accessed.

If your account is on one of he servers being serviced/maintained/upgraded, you’ll see a message like this:

Continue reading