Guest author Simon Freedman is product and partner marketing manager for Flexiant, a leading global provider of cloud management and orchestration software.
Say you want to deploy a multi-tier application on the cloud, but don’t know how much demand for it there will be at any given time. Without being able to predict traffic levels, how do you know how many servers each tier needs to be deployed on?
Choosing the right number of servers
If you run your application on too few servers, performance could suffer. If you run it on too many servers, you could be paying for infrastructure that you don’t need. To make things more complicated, given that demand will likely fluctuate over time, so too will the required amount of servers.
It’s simply not feasible to manually keep track of how the application is performing 24×7, and then spin up or kill off servers individually to cope with the change. What you need is the ability to automatically scale up and down as needed—which is exactly why Flexiant Concerto has integrated with New Relic.
New Relic APM’s best-of-breed monitoring is designed to give you fantastic insights into the various factors that affect your application’s performance, and the Apdex score it outputs helps provide an overall view. By combining New Relic with Flexiant Concerto, you can autoscale on the basis of infrastructure resource metrics such as CPU or RAM utilization. But while this can be of significant benefit, it may not fully represent whether the end user is likely to be satisfied with how the application is responding.
Apdex to the rescue
For this reason we have developed the capability within Flexiant Concerto to autoscale on the basis of the Apdex score. Apdex is one of the most accurate and reliable way of understanding how the application is performing—and therefore the end user’s experience.
Let’s say you built a server array to run the frontend tier of your application. You can set whatever Apdex score threshold you consider to be the minimum required (0.5, for example), and automatically scale out your servers when the Apdex score dips below this point. These additional servers are all built from the same template as the ones you started off with, and are all instantly added to the same load-balancing pool.
You can also define what Apdex score you consider to be acceptable (0.9, for example) and automatically scale down again once your app reaches that point. All of this happens around the clock, without requiring manual intervention. It’s true application-aware autoscaling.
The Flexiant Concerto user interface also gives you customizable graphical reporting on CPU, memory, and network traffic metrics, as well as Single Sign On to the New Relic control panel for more advanced functionality.Flexiant Concerto lets you build, configure, deploy, automate, and scale your applications in the cloud—fast. Check out www.concerto.io to find out more.
Take advantage of our promo offer and manage up to 50 servers for only $15 per month, saving you more than 70%. Simply enter the promo code FLEXNEWRELIC when signing up. And for more information on how New Relic and Flexiant work together, go to newrelic.com/flexiant.