{"id":25707,"date":"2026-10-05T12:07:42","date_gmt":"2026-10-05T12:07:42","guid":{"rendered":"https:\/\/www.examlabs.com\/certification\/?p=25707"},"modified":"2026-10-29T11:44:15","modified_gmt":"2026-10-29T11:44:15","slug":"microsoft-az-104-key-concepts-and-connections","status":"publish","type":"post","link":"https:\/\/www.examlabs.com\/certification\/microsoft-az-104-key-concepts-and-connections\/","title":{"rendered":"AZ-104: How Azure Controls Connect"},"content":{"rendered":"<p>Azure services interact through boundaries that are easy to confuse. A user might have permission to manage a storage account without permission to read its objects. A virtual machine can be running while its application is unavailable. A private endpoint can exist while the client resolves the wrong address. A healthy backup job may protect data that the organization still cannot restore quickly enough. These are not isolated Azure quirks; they reveal a small set of concepts repeated throughout administration.<\/p>\n<p><a href=\"https:\/\/www.examlabs.com\/az-104-exam-dumps\">Microsoft AZ-104 Exam<\/a> measures the ability to implement and maintain resources across identity, storage, compute, networking and monitoring. This article concentrates on the cross-service control principles behind that work, using a fictional public library application to make each distinction observable. It is not a repeat of the five-domain blueprint or another step-by-step exam plan. The aim is to identify which type of control can actually enforce the desired result.<\/p>\n<h2>Scope tells you where a management decision applies<\/h2>\n<p>Azure Resource Manager organizes management operations around resource groups, subscriptions and related resource hierarchies. Role-based access control defines which principal can perform which management actions at a scope. Azure Policy evaluates resource configuration against assigned requirements. These controls are complementary, but a role assignment and a policy assignment do not grant the same kind of authority.<\/p>\n<p>Imagine a local library deploying a public event website. A junior operator may be authorized to update a web application in one resource group. That permission should not imply authority to disable a subscription-wide security requirement or delete unrelated storage. A policy might require particular network settings even when the operator has enough role permissions to request the deployment.<\/p>\n<p>A valid deployment request may therefore be denied for policy reasons. Giving the operator a broader role may make the incident worse without satisfying policy. To diagnose the issue, identify the requested operation, principal, role, scope and policy evaluation separately. A good administrator can explain why each control exists and which one must change, if any.<\/p>\n<p><a href=\"https:\/\/www.examlabs.com\/certification\/an-introduction-to-azure-resource-manager\/\">Azure Resource Manager<\/a> helps explain how resource deployments and governance controls meet. The durable principle is that authority over a resource and permission for a specific configuration are different questions.<\/p>\n<h2>Authentication is not authorization, and authorization is not reachability<\/h2>\n<p>The library&#8217;s employees sign in with Microsoft Entra identities to manage event information. Authentication establishes the identity accepted by the relevant system. Authorization decides what that identity may do. Network controls determine whether a request can reach an endpoint at all. Treating these three gates as one \u201caccess\u201d problem produces unnecessary changes and misleading incident reports.<\/p>\n<p>Suppose the library website needs to display a protected image from Blob Storage. The application&#8217;s managed identity has a scoped read permission, but the storage account now accepts only private endpoint traffic. If the application&#8217;s DNS route still points to a public endpoint, correct authorization cannot solve the network failure. Similarly, allowing the path will not grant an identity permission to read a blob.<\/p>\n<p>Test each gate independently where possible. Confirm which principal is used, what operation is permitted, which name the client resolves, whether the selected route reaches the service and how the service evaluates the data request. A successful storage read is the combination of those conditions rather than evidence that all network or permission settings are generally correct.<\/p>\n<p>The same distinction applies to databases, management APIs and application records. A legitimate user who can reach a service may still be forbidden from reading another customer&#8217;s data. A network security group cannot enforce arbitrary business-record ownership if the application does not implement the check itself.<\/p>\n<h2>Desired configuration and effective behavior can disagree<\/h2>\n<p>Infrastructure templates and declarative tools describe what a deployment should create. The actual deployed state can differ because of policy, dependency failures, manual changes or runtime conditions. Even when the resource definition matches the intended configuration, the workload may behave differently from what the design assumed.<\/p>\n<p>Consider a virtual machine deployed successfully into an approved subnet. The configuration might specify a network interface and a service port, yet the application process could be stopped, the guest firewall could block traffic or the backend database could reject a connection. A successful resource deployment is not a complete test of the resulting application.<\/p>\n<p>Use repeatable post-deployment validation: compare effective settings against the intended state, test permitted traffic and denied traffic, inspect guest or service health and verify a small synthetic business operation. Record a rollback method before changes. This is more dependable than repeatedly applying a template and assuming the next deployment will correct an unrelated workload defect.<\/p>\n<h2>A network connection is a path, not a single setting<\/h2>\n<p>Virtual networks, subnets, peering, route tables, name resolution, security groups and private endpoints influence different stages of connectivity. An administrator should draw the path from source to destination before selecting a rule or diagnosing a timeout. Name resolution can point a client to an unintended address; routing can send the packet along the wrong path; filtering can deny an otherwise reachable service.<\/p>\n<p>The library runs a reservation API and a separate document archive. If a VM reaches one service but not the other, compare DNS, selected routes, allowed ports, endpoint exposure and application authorization. A broad \u201copen everything\u201d rule might mask the failing layer while creating unwanted access elsewhere.<\/p>\n<p>The <a href=\"https:\/\/www.examlabs.com\/certification\/step-by-step-guide-to-configuring-and-managing-virtual-networks\/\">Azure virtual networks<\/a> explanation is useful when paired with a source-and-destination test. A subnet&#8217;s existence does not prove that a workload has a correct route, and a correct route does not prove that the security rule or application permits the action.<\/p>\n<p>Network controls also have ownership boundaries. The library may manage Azure virtual-network settings but depend on a third-party firewall or ISP for part of a hybrid path. An Azure-only diagnostic result cannot prove that every external hop behaves correctly.<\/p>\n<h2>Redundancy protects a failure mode; it does not guarantee recovery<\/h2>\n<p>Availability, scaling, backups and replication solve different problems. Adding instances can improve capacity or availability under particular failures. Replication can provide another copy of data but may also replicate incorrect changes. Backups preserve recoverable points only if retention, permissions, keys and restoration procedures work together.<\/p>\n<p>If the library&#8217;s event site runs in multiple zones, a zone outage may not interrupt its web interface. But if all instances depend on one unavailable identity service, database or DNS configuration, the application may still fail. Conversely, a single-zone application with reliable backups may be recoverable but not meet the required downtime objective.<\/p>\n<p>State the recovery time objective and recovery point objective before choosing features. The library may tolerate an hour of event-listing interruption but almost no loss of paid reservation transactions. Those requirements can call for different strategies within the same overall service.<\/p>\n<p>Use a controlled restoration test and check data integrity and application access afterward. The <a href=\"https:\/\/www.examlabs.com\/certification\/unveiling-the-lesser-known-facets-of-azure-backup\/\">Azure Backup<\/a> mechanisms matter in that context. A green backup status is useful evidence about one operation, not definitive proof that the business can resume service after a disaster.<\/p>\n<h2>Storage security and cost are coupled through data behavior<\/h2>\n<p>Storage decisions affect not only how data is read but also durability, redundancy, access pattern, lifecycle and cost. An archive containing rarely viewed historic library records has different performance and recovery requirements from active circulation transactions. Choosing a storage tier solely because it is cheaper per gigabyte can create unacceptable retrieval delay or operational expense.<\/p>\n<p>Permissions should be scoped to the task. A read-only public event page does not need an identity capable of modifying all storage objects. Sensitive borrower records may need tighter access controls and explicit retention handling. Encryption at rest helps protect specified threats but does not replace correct application authorization after a privileged process decrypts content.<\/p>\n<p><a href=\"https:\/\/www.examlabs.com\/certification\/comprehensive-guide-to-azure-storage-solutions\/\">Azure storage<\/a> choices become easier when the application requirement is stated first: data structure, frequency of access, latency, recovery tolerance, privacy and operating cost. A feature checklist without those requirements is not an administrative design.<\/p>\n<h2>Monitoring should close the loop between expectation and evidence<\/h2>\n<p>The library&#8217;s website can show low CPU utilization while every reservation fails because its payment callback service is unavailable. Azure Monitor may report resource metrics, log events and alerts, but the administrator must select signals that answer the real operating question.<\/p>\n<p>Define a synthetic reservation that uses test data and follows the relevant dependencies. Determine which observation demonstrates success and which logs would distinguish authentication, application, network and payment-provider failure. Track the time and correlation identifiers without exposing customer details unnecessarily. An alert is useful when someone knows what action it informs.<\/p>\n<p>The <a href=\"https:\/\/www.examlabs.com\/certification\/what-is-azure-monitoring-a-complete-guide\/\">Azure Monitor<\/a> concepts apply across compute, storage and networking, but a metric needs interpretation. A missing record can indicate a collection problem or retention gap rather than proof that no incident occurred.<\/p>\n<h2>Use one test to verify multiple control boundaries<\/h2>\n<p>Choose a harmless event-page update: deploy a new image, read a permitted storage object, reject an unauthorized attempt, confirm that the website loads, inspect network behavior and verify that monitoring records the relevant transaction. Then change one condition at a time and note exactly which observation changes.<\/p>\n<p>When the permitted user fails but an unauthorized user succeeds, the issue is different from a total network outage. When a backup succeeds but a restore fails, the failure is not fixed by repeating the backup schedule. These paired positive and negative tests reveal which controls are doing their intended jobs and which assumptions remain unproven.<\/p>\n<p>Azure administration becomes more coherent when each concept is assigned to the boundary it governs: resource scope, identity, permission, route, workload configuration, resilience or evidence. Those distinctions are the foundation for diagnosing unfamiliar systems rather than memorizing a separate answer for every Azure service name.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Azure services interact through boundaries that are easy to confuse. A user might have permission to manage a storage account without permission to read its objects. A virtual machine can be running while its application is unavailable. A private endpoint can exist while the client resolves the wrong address. A healthy backup job may protect [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1648,1647],"tags":[],"class_list":["post-25707","post","type-post","status-publish","format-standard","hentry","category-all-certifications","category-certifications"],"_links":{"self":[{"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/posts\/25707","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/comments?post=25707"}],"version-history":[{"count":3,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/posts\/25707\/revisions"}],"predecessor-version":[{"id":28531,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/posts\/25707\/revisions\/28531"}],"wp:attachment":[{"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/media?parent=25707"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/categories?post=25707"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/tags?post=25707"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}