99.96% uptime and 600ms response time – the cloud really rocks (and so does mobile)

Data from February show that our 32 servers had an average uptime of 99.96%. As you imagine, hosting that many WordPress sites and keeping them all available 24×7 is no mean feat and this is where we rely heavily on the cloud. All our clients use the cloud to host their digital services and it really shows the scaleable, security and stability of cloud-based services.

In fact, just three sites were unavailable for more than 30 minutes and two of those were planned maintenance. Our “worst” server was for our internal accounting purposes and we lost it for over 2hrs (99.7% uptime) but this fell out of operating hours so wasn’t even missed.

WordPress support and outages in Feb

The average “period of unavailability” was less than 2 minutes in the whole month. One month has nearly 45,000 minutes so that’s quite a performance. Over a third of the sites had zero downtime or slow-downs. 107 outages is a little frustrating, but (again) our internal accounting system was 25% of those outages and a interconnect problem in Mexico was responsible for over 15 outages.

All servers on high-availability contract had zero downtime, as you’d hope for.

WordPress Hosting Performance

Critical to business success is the performance of the WordPress servers, how fast do they respond and how quickly is content served up. The average for February was a pretty impressive 600ms to reach and load content from our of our managed WordPress servers. We can thank extensive use of CDNs such as Amazon Cloudfront and NetDNA.

The future, our latest WordPress based installations are faster than the older ones – we’re moving in hyper-performance territory where the latest site released are regularly beating 300ms response time. This is ever more critical as mobile accesses to the cloud servers topped 1/3rd for the first time.