The Secret to Healthy App Delivery at Scale

by Iwan Price-Evans on Scalability • March 24, 2020 The Secret to Healthy App Delivery at Scale

When your IT operations or DevOps teams need to scale-up or scale-out the server systems running critical applications, it’s generally a good sign of business growth. It means there are more visitors coming to your website, more conversions, more product sales or possibly regional or international expansion. But with scaling comes complexity, especially for monitoring and managing application performance. This blog explains the basics of managing Application Delivery Control (ADCs) across multi-cloud deployments, multiple locations and at any scale.

ADCs are the foundation for ensuring your business-critical applications are always up and secure, by providing load balancing, web acceleration, web application firewall (WAF) and global server load balancing (GSLB).

When traditional hardware or software ADCs are scaled to tens, hundreds or thousands of instances, it becomes too complex and time-consuming to manage them manually and individually. Operators simply can’t implement new configuration or respond to problems fast enough. Also, a large-scale legacy ADC deployment is like a black hole in that it’s nearly impossible to see data about the health of all your servers and application performance in one place. Complexity and lack of visibility have been huge problems for many large businesses and organizations.

Today, the design and delivery of applications are changing radically with the increasing use of public and private clouds and cloud-native principles, such as containers, service meshes and microservices architectures. The implication for ADCs is that they must be cloud-native too. In addition to basic ADC functionality, the products also need to excel at East-West load balancing as well as enable multi-cloud and multi-location deployments.

And, most importantly, DevOps and IT Ops require visibility and insight into application metrics for all ADC nodes for any size of deployment – small-scale or hyperscale – and in any cloud environment. There are several features that today’s ADCs must have to deliver the required level of visibility.

Centralized, “Single Pane of Glass”

By moving configuration and management functions from the data plane path to the control plane, any number of ADCs can be controlled by a centralized cloud controller providing management, reporting, analytics and monitoring that your DevOps and IT Ops teams can access from an intuitive “single pane of glass” user interface.

This allows you to set configuration parameters once and launch them to all ADC nodes at the same time, removing the chore of having to configure each node individually. 

From a centralized controller, you can see a summary of all telemetry data and statistics from every ADC node, such as number of connections and requests as well as connection and request rates, data transfer and upstream health.

Real-Time, Granular Monitoring

Real-time monitoring is essential for ensuring application performance and securing your servers. Examples of just some of the questions you’ll want instant answers to:

  • Has anything changed?
  • What is the latency like?
  • Which pages are responding slowest?
  • Are there more failed logins from a specific country?

The centralized cloud controller should continuously monitor ADC nodes to provide real-time statistics, including response times (i.e., the time for waiting for a reply and the total time to respond to a client); connection rates for connecting upstream servers; session rates; percentage of idle time; number of cached items, size of cached objects and rate of cache hits; as well as attack statistics.

Customizable Reporting and Alerts

Everyone has preferences for what telemetry data they need to see and how it’s presented. The centralized cloud controller should allow you to personalize your view of the data and customize the reporting and alerts to your needs.

You might want live reporting on response times, for example, but you could also view response times statistics for the last two hours or last 24 hours. You determine how and when to view data on app performance, live stats, per IP rates and attack statistics. The platform responds to events in real-time, and you can set parameters for when alerts are raised and how you are notified.


When cloud-native ADCs are deployed and scaled, the real value will be in a centralized management platform. This is the brains of the operation that can provide the intelligence and insight you need to keep your business-critical apps always up and running, fast and secure.

A key feature of Snapt Nova is the centralized cloud controller, Nova Cloud. To see Nova Cloud in action and see the benefits of centralized control when scaling out ADCs, get started with the free Nova Community Edition.

Get the Snapt Nova Community Edition