Hosting Status Overview August 2011

Total uptime in August 2011 on Ad Delivery Platform 1: 99,931% (31 minutes downtime)

Total uptime in August 2011 on Ad Delivery Platform 2: 99,879% (54 minutes downtime)

Total uptime in August 2011 on Content Delivery Platform: 99,946% (24 minutes downtime)

In the night from August 17 towards August 18, around midnight CET, a number of short breaks occured in the core switches that connect our servers to the internet. Our supplier recommended to perform a software downgrade for these swithes, this was executed between midnight and 3 AM CET, resulting in a few more short network interruptions. These disruptions lasted on platform 1 for a total of 21 minutes, on platform 2 for a total of 24 minutes and on the content delivery platform for a total of 19 minutes. We took the opportunity of this non-planned emergency maintenance to implement a number of the new 20 Gbps core switches.

Op August 19, a short 1 minute disruption occured on both ad delivery platform 1 and the content delivery platform.

Op August 24, a short 1 minute disruption occured on both ad delivery platform 2.

On August 26 at 1.51PM CET, our primary load balancer was hit by hardware failure. After the network traffic was automatically transferred to the secondairy load balancer, problems started to emerge in the DNS function that was running on the same hardware at the time. Our technicians responded by setting up a dedicated DNS server, and also a redundant secondairy DNS server. When the primairy load balancer came back online at 2.48PM CET, full network capacity was available again. During the incident all platforms were slower than normal, and in addition a number of network interruptions occured which lasted a total of 9 minutes on platform 1, a total of 29 minutes on platform 2 and a total of 4 minutes on the content delivery platform.

On Tuesday August 2nd, shortly after noon CET, a serious hacking-incident occured. Even though this did not lead to hosting-downtime in the technical sense of the term, it caused major disturbances in the affected accounts. We published a full report shortly after the incident.