Concept questionMedium~7 min

Interview question

How would you trace one user request through an Azure-hosted .NET API and its downstream dependencies?

How would you trace a request across Azure-hosted services?

Explains correlation IDs, distributed tracing, dependency telemetry, message boundaries, and logs across App Service, queues, and downstream APIs.

1Say itAnswer out loud first

2CompareCompare with Core answer

3RefineRefine with Stronger and Advanced

TL;DR

I use W3C trace context so one operation can be followed across HTTP services and dependencies. Each span has its own identifier while sharing the trace identifier, and logs include those values automatically where instrumentation supports it. For asynchronous work I propagate trace context in message metadata and create a consumer span from it. Correlation values stay free of secrets and personal data because they spread through telemetry.

Full answer below

Practice focus

Correlation IDs, distributed tracing, trace context propagation, structured logs, dependency telemetry, queues, and investigation flow.

Answer first

Answer path

Core answer first, then checkpoints toward a stronger answer.

Start
Core answer
Start here. This is the direct spoken answer to practice first.
Why this question matters
When a request crosses APIs, databases, and queues, isolated logs are not enough. This drill matters because the production behavior is shaped by platform configuration, code boundaries, and the operational path the team will use during a real incident.
I use W3C trace context so one operation can be followed across HTTP services and dependencies. Each span has its own identifier while sharing the trace identifier, and logs include those values automatically where instrumentation supports it. For asynchronous work I propagate trace context in message metadata and create a consumer span from it. Correlation values stay free of secrets and personal data because they spread through telemetry.
Checkpoint
Stronger answer
Optional upgrade with examples, trade-offs, and sharper phrasing.
The hosting concern is reliable propagation. HTTP instrumentation carries traceparent; queue producers inject context and consumers extract it instead of inventing an unrelated id. Retries and fan-out create new spans that preserve their relationship to the original operation. Sampling and safe low-cardinality attributes keep the trace useful at production volume. Durable job, message, and idempotency identifiers remain separate because they answer business and delivery questions that a trace id does not.
Finish
Advanced answer
Optional deeper angle for production judgment and nuance.
I verify propagation across every library and asynchronous boundary in staging, including retries and broker redelivery. A long-running process may use a new trace linked to the initiating trace rather than pretending one trace lives forever. Azure Monitor or another backend should let an operator move from request to dependencies and worker activity without custom joins for every service. The goal is an explainable execution path, while durable identifiers preserve identity after trace sampling or retention has removed the original telemetry.

Review signals

Use these to catch weak phrasing, shallow reasoning, and risky claims.

Common mistakesCandidateCommon mistakesMistakes that weaken the answer or signal shallow understanding.Open details

Details

Generating a new unrelated ID in every service.
Not propagating context through queue messages.
Logging only free-form text without structured fields.
Putting sensitive data into correlation identifiers.

Red flagsCandidateRed flagsSignals that an interviewer may treat as risky.Open details

Details

Cannot explain why logs and traces need the same operation context.
No plan for async/background boundaries.
Debugs distributed failures by timestamp guessing only.
Does not mention dependency telemetry.

Follow-up prompts

Quick interviewer prompts inside this drill.

Follow-up questionsCandidateFollow-up questionsLikely next questions after the main answer.Open details

Details

Treat these as the next interviewer turn. Answer one out loud, then save the gap in your practice notes.

What fields belong in structured logs?
How do queues break normal request tracing?
How would you trace one failed order?
What data should never be used as a correlation ID?

Continue practice

No explicit flow link is set yet, so these are nearby drills from the same path or taxonomy.

Reversible App Service ReleaseArchitecture scenario / Medium / ~10 minContinue Deployment SlotsConcept question / Easy / ~5 minContinue App Service Plan ScalingConcept question / Medium / ~7 minContinue

Practice route

Read core answer Review checklist Follow-up prompts Track progress

Finished this drill?

Practice freely. Sign in only when you want to save completion, confidence, notes, or bookmarks.

New to Aporeon? Create account

Drill details

Share the question with a concise Core-answer excerpt and invite other developers to add their perspective.

Was this drill useful?

Help improve the interview library.

Found this answer useful?

Say thanks with a standalone, one-time $5 contribution. No account required.

No subscription or perks.Choose another amount

Human editorial direction

Aporeon is shaped by Aleksandar Tomovski, a software developer with experience on both sides of technical interviews. Content is reviewed for accuracy, natural spoken delivery, useful depth, and honest trade-offs.