Optimizing Azure Application Gateway with the Well-Architected Framework

Azure Application Gateway is a web traffic load balancer that works on Layer 7 of the OSI model and enables you to manage traffic for your web applications. It can make routing decisions based on attributes of an HTTP request such as URI path or host headers. The service also offers great app development features like autoscaling, SSL termination, support for Web Application Firewall, and the ability to dynamically modify request and response headers and request URL.

The Microsoft Azure Well-Architected Framework provides a set of architecture best practices to help you build and deliver great solutions. The framework is divided into five pillars of architectural best practices: cost management, operational excellence, performance efficiency, reliability, and security. These pillars help you effectively and consistently optimize your workloads against Azure best practices and the specific business priorities that are relevant to you or your customers' cloud journey. In addition to the general architectural guidance organized per-pillar in the Well-Architected Framework documentation, we are now publishing service-centered guidance in the form of service review guides that enable you to systematically validate and optimize individual components in your solution. The service review guides are concise and comprehensive guides with actionable recommendations for each Well-Architected pillar, derived from real-life experience. They include a checklist of design topics to pay attention to for each pillar, as well as configuration recommendations and Azure Policies you should enable.

Azure Well-Architected Framework review for Application Gateway v2 gives you simple and direct advice to optimize Azure Application Gateway implementations. It includes service-centered design checklists for each framework pillar. And recommendations that go from practical advice of things to take into consideration to avoid common pitfalls, like avoid removing backend servers too fast after making routing changes to give the service enough time to drain existing connections, to a formula you can use during capacity planning to estimate the number of instances needed to support a given load.

Jose Varela

Principal Software Engineer, Azure Architecture Center

Jose is an experienced software developer and solution architect with over 30 years of industry experience. Jose spent 20 years in various roles in Consulting and Architect roles in the Microsoft Services organization in North America before joining the Architecture Center, building custom solutions and providing architecture advice to Microsoft’s enterprise customers.