Stormagedden: Preparing for a Digital Deluge
By Aaron Rudger | December 11, 2014
As the West Coast braces for a powerful storm, public agencies and utilities are preparing their online services to keep important information and assistance flowing. Communities will look online—most likely on their wireless devices—to websites, mobile apps, and email to see if roads are clear, classes have been cancelled, or when their local electrical outage will be restored. Events like this are reminders of the need for performance testing and synthetic monitoring.
Hopefully, the utility service your neighborhood has tested their websites or apps with tests that truly validate good performance under peak traffic. The only way to know this is to simulate the expected number of users and their journeys through the site with real Internet traffic in production. For example, a utility worker having to search for customer data during a power outage and maps to show utility generators.
Production environments are larger scale, and depend on different external systems compared to a staging or test environment. Testing services that use transactions generated from cloud computing technologies can replicate the conditions that actual users will experience in production. This type of load testing models customer behavior, and the ways in which users continually arrive on a site during peak events, no matter if on a laptop, smartphone or tablet. These tests simulate and measure the total user journey such as a typical path from login to looking up a customer address, maybe nearby stations, and what requests or searches are needed to troubleshoot and then resolve problems.
Thorough web load testing will result in errors and bottlenecks—maybe even a full-scale crash—and that is the desired result. The objective is to identify issues with garbage collection, caching, database connectivity, among others. This step also allows for collecting a large amount of data. If tested correctly, an end-to-end transaction performance analysis provides enormous insight into network, infrastructure, and application performance.
Regardless of the testing an agency or utility may (or may not) have made, continuous monitoring—especially during a storm event, for example—is critical. Proactive testing from a service like Keynote quickly alerts operations teams to errors and scalability issues. With diagnostic detail about performance and availability as experienced by customers using desktop and mobile devices, issues can be resolved quickly and customers can get the information they need.
Despite the best testing, no system is perfect. Problems happen, sometimes out of the control of a site/app owner. Just like it’s a good idea to keep extra matches, flashlights and water on hand, public sector and utility sites can turn to services from cloud providers like Keynote to help them weather the storm.