{"id":4507,"date":"2026-10-12T21:32:57","date_gmt":"2026-10-12T21:32:57","guid":{"rendered":"https:\/\/www.examtopics.info\/blog\/?p=4507"},"modified":"2026-10-12T21:32:57","modified_gmt":"2026-10-12T21:32:57","slug":"microsoft-az-305-front-door-app-gateway-global-routing","status":"publish","type":"post","link":"https:\/\/www.examtopics.info\/blog\/microsoft-az-305-front-door-app-gateway-global-routing\/","title":{"rendered":"Microsoft AZ-305: Front Door, App Gateway, and Traffic Manager"},"content":{"rendered":"<p>A global ticketing website receives requests from customers in Europe, North America, and Asia. Its order API lives in two Azure Regions, but payments can only be written to one authoritative regional database at a time. During a regional incident, DNS should direct users away from a failed site without turning a payment retry into a second order. An architect is asked to choose between Azure Front Door, Application Gateway, Load Balancer and Traffic Manager. The correct design depends on protocol, routing scope, inspection, TLS, and application state\u2014not on treating every load-balancing service as the same box.<\/p>\n<p><strong>Microsoft AZ-305<\/strong> includes choosing an internet connectivity, routing, load-balancing and application protection architecture. The existing <a href=\"https:\/\/www.examtopics.info\/blog\/microsoft-az-305-highly-available-azure-applications\/\">highly available Azure applications<\/a> specialist covers availability as an end-to-end property. This article narrows the question to entry points: which component decides where a request goes, which layer inspects it, and how a client recovers when a healthy-looking origin cannot complete the business operation.<\/p>\n<h2>Decide at which layer routing must understand the request<\/h2>\n<p>Layer 4 load balancing makes decisions around transport connections, while layer 7 application delivery can use HTTP hostnames, paths, headers or other supported request attributes. DNS-based routing chooses which address a resolver returns but cannot redirect an already-established connection instantly. An architect should describe the protocol first: HTTPS web pages, UDP gaming traffic, a long-lived TCP application, and a private API are different ingress requirements.<\/p>\n<p>Write the user&#8217;s route as a series: DNS or global edge, TLS handshake, security inspection, regional gateway, application compute, and data tier. Then mark where user identity is checked and where a write becomes irreversible. This separates a routing failure from an application consistency problem that no load balancer can solve on its own.<\/p>\n<h2>Use Azure Front Door for supported global HTTP delivery<\/h2>\n<p>Azure Front Door is an edge-based HTTP(S) application delivery service that can optimize global routing, caching, and supported security policies. Current Standard and Premium capabilities vary, so inspect the tier when a requirement needs WAF controls, Private Link to origins or other features. The <a href=\"https:\/\/learn.microsoft.com\/en-us\/azure\/frontdoor\/front-door-overview\">Microsoft Front Door overview<\/a> describes the supported edge model; do not assume a classic-tier option is the right baseline for a new design.<\/p>\n<p>A multinational website may use Front Door to direct users toward healthy regional origins and cache public static assets. Keep personalized or regulated responses out of shared caching unless the cache key and authorization model explicitly make reuse safe. If a Region loses its payment database, a successful origin HTTP response from a lightweight liveness check might be misleading; origin health must represent real ability to serve the intended request path.<\/p>\n<h2>Use Application Gateway when regional HTTP inspection matters<\/h2>\n<p>Azure Application Gateway provides regional, application-layer traffic routing for supported web workloads. It can support host- and path-based rules, TLS handling, and a Web Application Firewall tier where selected. This is useful when regional applications need an HTTP ingress boundary, backend pools, and request-aware routing inside an Azure network. A global edge and a regional gateway may work together, but every additional hop adds certificate, header forwarding, logging and failure responsibilities.<\/p>\n<p>Design the health probes around readiness. A backend that accepts TCP port 443 may be unable to query its database. Conversely, a health probe that requires every optional analytics service to succeed can mark all targets unhealthy during a minor reporting outage. Define a narrow essential transaction for routing decisions, not a generic \u201cevery dependency up\u201d endpoint.<\/p>\n<h2>Use Azure Load Balancer for transport-level scenarios<\/h2>\n<p>Azure Load Balancer works at the transport layer and serves supported public or internal traffic scenarios. It is appropriate when the application needs network-level connection distribution rather than HTTP path routing. Private services may use an internal load balancer for workload communication behind controlled network security groups. It cannot replace a regional layer-7 gateway when URL routing or HTTP WAF inspection is central to the requirement.<\/p>\n<p>Architecture reviews should also distinguish availability zones, backend health probes, session persistence, and cross-zone traffic behavior. A load balancer with two backend instances in one Availability Zone is not a full multi-zone design. The <a href=\"https:\/\/www.examtopics.info\/blog\/microsoft-az-305-hub-and-spoke-networking-in-azure\/\">hub-and-spoke network analysis<\/a> helps locate internal routing and inspection services so transport routing reinforces rather than bypasses segmentation.<\/p>\n<h2>Understand Traffic Manager&#8217;s DNS steering limits<\/h2>\n<p>Azure Traffic Manager uses DNS-based traffic distribution for supported routing methods, such as priority or performance, using monitored endpoints where configured. It does not proxy each HTTP request, terminate TLS for the app, or make clients forget cached DNS answers immediately. Use it when DNS-level selection is the right level of control, and understand that resolver caching and application connection lifetimes affect recovery time.<\/p>\n<p>If a company needs request-level edge protection, offload and routing, compare Front Door rather than imagining Traffic Manager has identical capabilities. A DNS-based failover may still be valuable for non-HTTP endpoints or architectures where a proxy is unsuitable. Write the recovery expectation with realistic DNS behavior rather than promising instant global failover from an aggressive TTL setting.<\/p>\n<h2>Put WAF where attackers can reach the application<\/h2>\n<p>A WAF helps screen selected HTTP request patterns. It is not a substitute for application authorization, tenant access controls, or backend database protection. If the WAF protects Front Door but the origin can also be reached over an unrestricted public endpoint, the intended security boundary may be bypassed. Restrict direct-origin paths with supported private connectivity or resource and network policies when the design requires it.<\/p>\n<p>Model legitimate high-volume traffic as well as malicious patterns. A sale launch can look like a denial-of-service event to a simplistic rate rule. Start with observable policies and exceptions that have owners and expiry conditions. Retain routing and security logs that allow investigators to distinguish a WAF block from a load-balancer health rejection or an application-generated authorization error.<\/p>\n<h2>Design multi-region failover around the data owner<\/h2>\n<p>A global front door can route requests to a second region, but that region must possess enough compute and an authorized, consistent data path to complete the operation. If only the primary region can write orders, simply sending POST requests to the secondary may cause lost or duplicate transactions. Define read-only degradation, queueing, database promotion, or an explicit authoritative-region transition as business behaviors before changing routing weights.<\/p>\n<p>The <a href=\"https:\/\/www.examtopics.info\/blog\/microsoft-az-305-azure-disaster-recovery-with-paired-regions\/\">paired-region disaster-recovery article<\/a> establishes geographic recovery concepts. For application delivery, pair its RTO\/RPO assumptions with realistic probe frequency, endpoint capacity, DNS or edge route convergence, credential availability and safe recovery of session state. A failover that moves users but not write authority has not restored the application.<\/p>\n<h2>Test origin failure and edge policy independently<\/h2>\n<p>In a controlled environment, make one regional application origin return HTTP 503 while the other remains fully healthy. Predict which component detects the failure and how long customers continue using the affected route. Then test a direct-origin request that should be blocked and a legitimate HTTP request that should be allowed through WAF. Capture client-visible status, backend health, logs and failover timing.<\/p>\n<p>Finally, estimate cost from global edge requests and transfer, regional gateways, WAF processing, idle redundancy and monitoring. A simple internal application may not need a global edge tier, while a large public service may justify it to improve latency and protection. AZ-305 architects should choose an ingress stack whose routing decisions, failure semantics, security controls and costs are understandable before the first outage.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Choose global and regional ingress from protocol, edge security and failover semantics.<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[10,1],"tags":[],"class_list":["post-4507","post","type-post","status-publish","format-standard","hentry","category-networking","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/www.examtopics.info\/blog\/wp-json\/wp\/v2\/posts\/4507","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.examtopics.info\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.examtopics.info\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.examtopics.info\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.examtopics.info\/blog\/wp-json\/wp\/v2\/comments?post=4507"}],"version-history":[{"count":1,"href":"https:\/\/www.examtopics.info\/blog\/wp-json\/wp\/v2\/posts\/4507\/revisions"}],"predecessor-version":[{"id":4519,"href":"https:\/\/www.examtopics.info\/blog\/wp-json\/wp\/v2\/posts\/4507\/revisions\/4519"}],"wp:attachment":[{"href":"https:\/\/www.examtopics.info\/blog\/wp-json\/wp\/v2\/media?parent=4507"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.examtopics.info\/blog\/wp-json\/wp\/v2\/categories?post=4507"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.examtopics.info\/blog\/wp-json\/wp\/v2\/tags?post=4507"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}