Architecture scenarioHard~12 minApplication Architecture Essentials / 11 of 15

Interview question

When would you use CQRS in a .NET application, and what does it not require?

CQRS: when separate read and write models help

Applies CQRS proportionally, separating model responsibilities without assuming separate databases, messaging, or event sourcing.

1Say itAnswer out loud first

2CompareCompare with Core answer

3RefineRefine with Stronger and Advanced

TL;DR

I consider CQRS when writes enforce rich invariants but reads need very different shapes, joins, filtering, or scaling. Commands express intent and return an outcome, while queries return purpose-built read models without mutating state. In a modest application both sides can share one database and transaction technology. That lightweight separation can improve clarity without introducing distributed consistency.

Full answer below

Practice focus

Command-query separation, read models, write invariants, scaling, consistency, complexity, and event-sourcing misconceptions.

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
CQRS separates models for changing state from models for reading it. It does not automatically require separate services, databases, a message bus, or event sourcing.
I consider CQRS when writes enforce rich invariants but reads need very different shapes, joins, filtering, or scaling. Commands express intent and return an outcome, while queries return purpose-built read models without mutating state. In a modest application both sides can share one database and transaction technology. That lightweight separation can improve clarity without introducing distributed consistency.
Checkpoint
Stronger answer
Optional upgrade with examples, trade-offs, and sharper phrasing.
The trade-off appears when read data is projected asynchronously or owned separately. Then clients may see stale results, projection failures need repair, and operations need correlation. I would not add that machinery to simple CRUD. Direct EF Core projections may be enough for reads, while write handlers use domain behavior and focused persistence. Command and query types should describe business operations rather than create a class-per-method framework ritual.
Finish
Advanced answer
Optional deeper angle for production judgment and nuance.
In production I define the consistency promise for each user flow. After a successful command, the response may include the new state, the client may poll an operation, or the read model may support read-your-own-write through a version. Projection consumers need idempotency, checkpoints, replay, lag metrics, and a repair path. Authorization must apply on both command and query sides. I would adopt CQRS in stages: first separate code paths, then separate storage or processing only when measured load, ownership, or model conflict pays for the operational cost.

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

Saying CQRS requires event sourcing.
Creating command and query classes for trivial CRUD with no benefit.
Ignoring stale reads after asynchronous projection.
Applying authorization only to commands.

Red flagsCandidateSigns the design is fragileSignals that an interviewer may treat as risky.Open details

Details

The product has no defined consistency behavior.
Projection lag is invisible.
Read and write models share hidden mutation.
A framework pipeline is more complex than the business operation.

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.

Can CQRS use one database?
How would you provide read-your-own-write?
When does a separate read store become worthwhile?

Continue the section

Next in Application Architecture Essentials: Modular monolith boundaries

Step 12 of 15

Enforce boundaries inside a modular monolith

Opening drill

Architecture scenarioHard~12 minApplication Architecture Essentials / 11 of 15

Interview question

When would you use CQRS in a .NET application, and what does it not require?

CQRS: when separate read and write models help

Applies CQRS proportionally, separating model responsibilities without assuming separate databases, messaging, or event sourcing.

1Say itAnswer out loud first

2CompareCompare with Core answer

3RefineRefine with Stronger and Advanced

TL;DR

Full answer below

Practice focus

Command-query separation, read models, write invariants, scaling, consistency, complexity, and event-sourcing misconceptions.

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
CQRS separates models for changing state from models for reading it. It does not automatically require separate services, databases, a message bus, or event sourcing.
I consider CQRS when writes enforce rich invariants but reads need very different shapes, joins, filtering, or scaling. Commands express intent and return an outcome, while queries return purpose-built read models without mutating state. In a modest application both sides can share one database and transaction technology. That lightweight separation can improve clarity without introducing distributed consistency.
Checkpoint
Stronger answer
Optional upgrade with examples, trade-offs, and sharper phrasing.
The trade-off appears when read data is projected asynchronously or owned separately. Then clients may see stale results, projection failures need repair, and operations need correlation. I would not add that machinery to simple CRUD. Direct EF Core projections may be enough for reads, while write handlers use domain behavior and focused persistence. Command and query types should describe business operations rather than create a class-per-method framework ritual.
Finish
Advanced answer
Optional deeper angle for production judgment and nuance.
In production I define the consistency promise for each user flow. After a successful command, the response may include the new state, the client may poll an operation, or the read model may support read-your-own-write through a version. Projection consumers need idempotency, checkpoints, replay, lag metrics, and a repair path. Authorization must apply on both command and query sides. I would adopt CQRS in stages: first separate code paths, then separate storage or processing only when measured load, ownership, or model conflict pays for the operational cost.

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

Saying CQRS requires event sourcing.
Creating command and query classes for trivial CRUD with no benefit.
Ignoring stale reads after asynchronous projection.
Applying authorization only to commands.

Red flagsCandidateSigns the design is fragileSignals that an interviewer may treat as risky.Open details

Details

The product has no defined consistency behavior.
Projection lag is invisible.
Read and write models share hidden mutation.
A framework pipeline is more complex than the business operation.

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.

Can CQRS use one database?
How would you provide read-your-own-write?
When does a separate read store become worthwhile?

Continue the section

Next in Application Architecture Essentials: Modular monolith boundaries

Step 12 of 15

Enforce boundaries inside a modular monolith

When would you use CQRS in a .NET application, and what does it not require?

Practice focus

Answer path

Why this question matters

Review signals

Follow-up prompts

Continue the section

When would you use CQRS in a .NET application, and what does it not require?

Practice focus

Answer path

Why this question matters

Review signals

Follow-up prompts

Continue the section