AZ-400 - Implement an Instrumentation Strategy - Section 5.2

Analyze metrics from instrumentation, including infrastructure performance indicators, distributed tracing, and Kusto Query Language (KQL) queries.

Analyse Application Insights telemetry using Kusto Query Language (KQL) to correlate distributed tracing data, performance counters, and custom events across services. Distinguish the infrastructure performance indicators and application-level signals that reveal bottlenecks and inform decisions about scaling or remediation.

distributed tracingKusto Query LanguageApplication Insightsperformance counterstelemetry analysis

Practice question for this objective

Free sampleImplement an Instrumentation Strategyhard

Three separate "Application Insights" resources, one per microservice, each hold their own requests telemetry. An engineer must run a single Kusto Query Language query that aggregates failure counts across all three resources at once, without first copying their data into one shared resource. Which approach lets one query span the separate resources?

  • AAdd a workspace-based diagnostic setting to each resource so all three forward telemetry into one Log Analytics workspace before any query is written.
  • BUse the join operator on operation_Id to combine the requests tables from the three resources into a single correlated result set for aggregation.
  • CApply the make-series operator over the requests table so the failures are expanded into a continuous time series across the chosen interval.
  • DReference each resource with the app expression inside a union, so one query reads the requests telemetry from all three resources and aggregates the failures together. Correct
The app expression combined with union lets one Kusto query read across multiple Application Insights resources without consolidating their data first. Cross-resource queries use the app function to address another Application Insights resource by name, and placing several such references inside a union merges their rows into one stream, so a single statement aggregates failures across all three resources while their telemetry stays in place.

Why A is wrong: Routing the resources to one workspace would centralise the data and is a sound long-term design, but it changes the ingestion configuration rather than letting a single query span the existing separate resources as the requirement states.

Why B is wrong: The join operator correlates rows by a shared key and is tempting because the query spans multiple inputs, but join matches records within reachable tables and does not by itself address telemetry held in other Application Insights resources.

Why C is wrong: The make-series operator builds a gap-filled time series and is plausible for trend analysis, but it shapes data from a single accessible scope and does nothing to reach across three distinct resources.

Why D is correct: The app expression names another Application Insights resource within a query, and combining several app references under a union lets one statement read and aggregate telemetry across all three resources at once.

See more AZ-400 practice questions, answers explained.

More in this domain

Back to all Implement an Instrumentation Strategy objectives, or the AZ-400 cert hub.

Examworthy is not affiliated with or endorsed by Microsoft. Original, blueprint-aligned practice material only.