Implement an instrumentation strategy

Your team runs an ASP.NET Core API instrumented with Application Insights. Support reports that certain checkout requests fail intermittently, but the front-end only logs a generic 500 error. You need a KQL query in Log Analytics that returns each failed checkout request together with the specific exception thrown during that same request execution, so you can identify the root-cause exception type per failing operation. Which query correctly correlates the two telemetry types?

Your team runs an Application Insights-instrumented web API. Product management wants a weekly report showing the 99th percentile server response time grouped by each API operation (URL path), covering the last 7 days. You need to write a KQL query against the requests table that returns one row per operation with its p99 duration. Which query satisfies this requirement?

Your team runs a high-traffic ASP.NET Core API instrumented with Application Insights. Finance flags that ingestion costs have spiked, and you want to reduce telemetry volume. However, the on-call engineers report that during incident investigations, related request, dependency, and exception records for the same operation are sometimes missing, making distributed traces incomplete. You need to cut ingestion volume while preserving statistically accurate metrics and keeping correlated telemetry items together for a given operation. Which approach best meets these requirements?

Your team runs a customer-facing web API on Azure App Service. Leadership wants proactive detection of outages from multiple global locations, including verification that a login endpoint returns a valid token before the checkout endpoint is exercised, all without deploying any code into the application. You are configuring Application Insights. Which monitoring capability should you implement?

Your microservices application is instrumented with Application Insights across five services. A customer reports that a specific checkout request failed intermittently. You have the operation_Id from the failed request captured in a support ticket. You need to reconstruct the complete end-to-end distributed trace — including all requests, dependencies, and exceptions across every service — for that single transaction using KQL. Which query approach correctly retrieves the full trace?

Your team runs a microservice on Azure App Service instrumented with Application Insights. After a recent release, users report intermittent slowness. You suspect a downstream SQL dependency is the cause. You need a KQL query that returns, for the last 24 hours, the 95th percentile duration of failed dependency calls grouped by the target dependency, so you can identify which dependency degraded most. Which query should you run in the Logs (Analytics) blade?

Your team uses Application Insights with Log Analytics. The reliability lead wants a dashboard tile showing the trend of the 95th percentile server response time bucketed into 15-minute intervals over the last 24 hours, visualized as a line chart. Which KQL query produces the correct result?

A production ASP.NET Core web app is instrumented with Application Insights. During a promotional event, on-call engineers report that the app is intermittently returning 500 errors, and they need to observe request rates, failure rates, and server CPU in real time as the traffic surges — with sub-second latency and without waiting for telemetry to be ingested and indexed. Which Application Insights capability should they use?

Your team runs a customer-facing web API instrumented with Application Insights. Leadership wants to be automatically notified when the failure rate of the API rises abnormally compared to its typical baseline, without manually defining static numeric thresholds that would need constant tuning as traffic patterns change throughout the day. You must configure this with the least ongoing maintenance. What should you use?

Your team runs a customer-facing API on an Azure Virtual Machine Scale Set. Traffic is highly seasonal and follows a predictable daily pattern that shifts gradually over weeks, so a fixed CPU alert threshold either fires constantly during normal peaks or misses genuine spikes during quiet periods. You must configure an Azure Monitor metric alert that automatically adapts to the changing baseline of CPU utilization while still catching abnormal deviations, and it must require minimal ongoing manual tuning. Which alert configuration should you use?

Your team runs a production workload on an Azure Kubernetes Service (AKS) cluster. You have enabled Container Insights, and users are reporting intermittent pod evictions. You need to determine whether the evictions correlate with node-level memory pressure across the cluster over the past 24 hours. Which action provides this insight most directly?

Your team stores application and infrastructure logs in a single Azure Monitor Log Analytics workspace. Compliance requires that a specific table, 'SecurityEvent', be retained and queryable for 30 days at interactive speed, but must remain retrievable for a total of 2 years to satisfy audit requests. Interactive retention across all tables is currently set to 90 days, which is inflating costs. You want to minimize storage costs while meeting the requirement with the least operational overhead. What should you configure?

Your team runs 40 Azure VMs across two subscriptions. Operations needs a single view showing per-VM CPU, memory, and disk performance plus a visual map of process-level TCP connections between the VMs to troubleshoot latency between an app tier and a database tier. You must enable this with the least ongoing agent management overhead and ensure telemetry lands in a specific Log Analytics workspace. What should you configure?