AZ-900 Describe Azure architecture and services Practice Question

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Azure Monitor Metrics

  Why it's wrong here

  Azure Monitor Metrics aggregates telemetry into numeric performance counters such as CPU utilization, memory consumption, and request counts, giving high-level views of resource health. Because it samples and summarizes data over time intervals, it lacks the operation-level correlation identifiers needed to trace a single user request through each dependency. Consequently, while it can alert on aggregate latency changes, it cannot show the step-by-step path that a specific request takes across services.

• ✓

  Azure Application Insights distributed tracing

  Why this is correct

  Azure Application Insights distributed tracing automatically assigns a unique operation ID to each incoming request and propagates that ID across outbound calls to services, queues, and databases. The resulting telemetry is correlated in the end-to-end transaction view, which visually maps every dependency, duration, and failure point along the request's path. This capability makes it the only option listed that directly and automatically reveals the full path of individual requests across multiple services.

• ✗

  Azure Log Analytics queries

  Why it's wrong here

  Azure Log Analytics queries give you raw log access and the ability to join tables using KQL, but they do not create correlation links on their own. Without trace context embedded in each log entry—typically an operation ID from Application Insights—you would have to manually infer relationships using timestamps and custom fields, which is error-prone and incomplete. Thus Log Analytics can assist in investigating known issues but cannot automatically present the full end-to-end path of a request across distributed components.

• ✗

  Azure Network Watcher

  Why it's wrong here

  Azure Network Watcher is designed for infrastructure-level network troubleshooting, offering tools like connection monitor, IP flow verify, and packet capture that inspect L3/L4 connectivity, routing, and NSG rules. It operates on IP packets and network interfaces rather than application-level HTTP requests or database calls, so it cannot observe the logical flow of an application request from front end to back end. As a result, it is fundamentally unable to reconstruct an individual request's distributed trace across services.