Debugging, Website Performance Issues & Real-Time Statistics – Hugo Cruz
How to use real-time statistics to pinpoint website performance issues and enhance user experiences There are two big problems that make visitors leave a website: slow loading times and irrelevant content. Failure to address those fast can impact a business negatively.Ten years ago, Amazon discovered during A/B tests that every 100ms of latency cost them
How to use real-time statistics to pinpoint website performance issues and enhance user experiences
There are two big problems that make visitors leave a website: slow loading times and irrelevant content. Failure to address those fast can impact a business negatively.Ten years ago, Amazon discovered during A/B tests that every 100ms of latency cost them one percent in sales. If anything, people have become much more impatient since then. However, websites are now so complex that it is often hard to pinpoint which part of the web architecture or interactions with it are causing performance issues. The same is true for quickly finding out which content turns users off or on. As with so many questions nowadays, the answer is: data or more specifically -real-time website traffic data. And there’s a perfect place to collect this data in real-time: Varnish Cache.
Varnish monitors and logs all traffic flowing through websites with almost no latency, so website owners can gather a host of information as it happens. But how to explore this data? The easiest way is to use Varnish Custom Statistics (VCS), a real-time statistics engine, that allows users to aggregate, display and analyse web traffic.
VCS can count everything: page popularity, click count, trending content, articles or sale products, where traffic is coming from, URLs with long load times, URLs with most hits, URLs with the most 50x response code errors, which clients are generating load, etc. VCS outputs data via a time series API in JSON format which can easily be integrated into existing monitoring solutions.
Norwegian, Møller Group is a good example of how using VCS’s JSON output data in combination with a monitoring tool can create real business value. Before using VCS, the company experienced load problems in its main database but did not know where the usage was coming from. On the API side, often the cause of load, they couldn’t visualise the data traffic. By integrating VCS with DataDog, Møller Group got deep insight into their performance problems and was able to pinpoint the cause of poor response times, the level and distribution of errors and to act quickly and resolve the problems.
In his practical session Hugo Cruz, Product Manager at Varnish Software, will talk about how to use VCS to gather web data in real-time and deliver better website user experience that makes visitors stay. He will introduce delegates to VCS, how to use itfor identifying performance issues and for more advanced use cases such as A/B Testing, tracking conversion rates, measuring click through-rates and counting users per video stream.