AZ-400 ยท Analyzing telemetry
Distributed tracing with Application Insights
Application Map, transaction search and end-to-end transaction view, operation IDs and correlation, OpenTelemetry, sampling and its effect on analysis.
Key points
- Items from one request share an operation ID and point to a parent, forming a tree across services.
- Application Map: components by cloud role name, HTTP calls between them, health indicators.
- End-to-end transaction view: timeline of one operation.
- Instrumentation: OpenTelemetry-based distro, W3C trace context.
- Sampling keeps or drops whole operations; per-service sampling can leave end-to-end views incomplete.
Source: https://learn.microsoft.com/en-us/azure/azure-monitor/app/app-map
Practice questions
Which field groups all telemetry that belongs to one end-to-end operation across services?
- cloud_RoleName
- session_Id
- operation_Id
- client_IP
Show answer
Answer: C. operation_Id
operation_Id is shared by the request, its dependencies, traces and exceptions. Role name identifies a component, session ID groups a user's activity, and client IP identifies a caller, none of which define one operation.
Which statements about Application Map are true?
- Nodes are identified by the cloud role name in telemetry
- Edges are derived from HTTP dependency calls between components
- Each component must live in a different subscription
- Components can be roles inside one Application Insights resource
- Only failed calls are ever drawn on the map
Show answer
Answer: A. Nodes are identified by the cloud role name in telemetry, B. Edges are derived from HTTP dependency calls between components, D. Components can be roles inside one Application Insights resource
Nodes are identified by cloud role name, edges come from HTTP dependency calls between instrumented components, and components may be separate resources or roles within one resource. A map is not limited to failed calls, and subscriptions do not have to differ.
True or false: ingestion sampling still discards items at the Application Insights endpoint when the SDK is already sampling that telemetry.
Show answer
Answer: False
False. Ingestion sampling only applies when no other sampling is in effect; if the SDK samples (adaptive or fixed-rate), the service ignores the ingestion rate for that telemetry.
Practice 8 questions on this topic in Bitzu
Short lessons, instant explanations and spaced review of what you get wrong. Free on iOS and Android.
