On June 24, several important internet services experienced outages. Citrix Intelligent Traffic Management (ITM) Radar observed another outage on July 2, this time on the Cloudflare CDN. Thousands of popular websites were affected by this brief outage, which lasted for about 30 minutes (1:45 p.m. UTC until about 2:15 p.m. UTC) as detected by Citrix ITM Radar and as shown in the Cloudflare availability chart below.

Citrix ITM can also be used to report on the platform availability specific to different countries, regions, states and networks. In this case, the outage appears to have had a worldwide impact. The chart below shows that users from several countries were unable to access services hosted on the Cloudflare CDN.

Citrix ITM Radar is the world’s largest internet performance experience measurement network, benchmarking the relative availability and performance of more than 35 global CDNs and hundreds of public cloud and hosting data centers.

ITM Radar collects real-time experience data from over 50,000 ISP networks worldwide, including over 15 billion daily data points from about 900 million end-user sessions. These data empower Citrix ITM to ensure predictive application delivery, confirming that the correct path is chosen based on real-time internet user data points.

What Can Be Done?

Citrix ITM enables you to avoid the problems of CDN and cloud outages. Citrix ITM customers can set their traffic management rules so that user requests are guided to available and reachable content based on the Radar data in real time. Users’ requests for content will be sent only to data centers, cloud regions, or CDNs that are available. Users will be blissfully unaware that an outage has occurred.

To illustrate the point, here’s the experience of an actual Citrix ITM customer. Like many companies, this customer hosts their applications on CDNs. Fortunately, the customer positions content across multiple CDNs and uses Citrix ITM to load balance between them.

Cloudflare Availability Chart

Citrix ITM Decision Chart

As you can see, when Cloudflare had normal availability, Citrix ITM distributed traffic to Cloudflare CDN along with the other CDNs the customer is using (point 1 in Cloudflare availability chart and point 4 in Citrix ITM decision chart).

When Citrix ITM noticed that Cloudflare was unavailable starting 1:45 p.m. UTC (point 2 in Cloudflare availability chart), it dynamically steered traffic to other available CDNs (between points 5 and 6 in the Citrix ITM decision chart). When the Cloudflare CDN came back online (point 3 in the Cloudflare availability chart), it was once again added to the traffic distribution algorithm and requests were sent to Cloudflare (point 6 in the Citrix ITM decision chart) whenever it was most optimal to do so.

All of this detection and recovery happened automatically. The customer didn’t have to react to  pager notifications or reconfigure settings for a failure scenario. Radar measured the issue from end users, and Citrix ITM automatically made the best choices to ensure the customer’s application was still able to respond to user requests. There was no impact on the ability for the service to deliver content and earn revenue.