System Design · Engineering Foundry challenge level: Advanced

Multi-Region Feature Flag Control Plane

Design a control plane and read path that distribute flag changes globally without making application requests depend on a central service.

All challengesSuggested practice time: 80 minutes

Original prompt

What you are building

Design a multi-region feature flag platform for internal services. Operators create versioned rules, applications evaluate flags locally, and emergency disables should propagate quickly. Produce an architecture that makes consistency, safety, rollback, tenancy, and regional failure behavior explicit.

Engineering Foundry originalNo company association

Deliverables

Assumptions and service-level objectives.
Control-plane and data-plane architecture.
Rule/version data model and distribution events.
APIs or configuration interfaces.
Reliability, rollback, and observability strategy.
Key tradeoffs and unresolved risks.

Constraints

Application evaluation must not require a network call on every request.
Rules must be auditable and tenant-isolated.
A bad rule must be recoverable without editing it in place.
Regional disconnection must have an explicit behavior.

Success criteria

Control and evaluation paths are separated.
Versioning and rollback behavior are concrete.
Propagation and stale-read risks are observable.
The design states how it behaves during regional and distribution failures.

Suggested workflow

Define ordinary and emergency change objectives.
Separate operator writes from application evaluation.
Model immutable versions and activation state.
Trace rollout, rollback, and regional partition scenarios.
Add observability and operational controls.

Attempt before reveal

Build your deliverable independently.

Use the prompt, constraints, and workflow first. Guidance offers considerations and questions - not a canonical solution.

Reveal guidanceDirections and questions to test

One direction to explore

Treat published rule versions as immutable artifacts and activation as a separate decision.
Compare streaming distribution with periodic snapshot recovery.
Ask which component can safely initiate an emergency disable when the primary control region is unavailable.

Reveal common mistakesFailure patterns worth checking

Putting a synchronous central lookup on the request path.
Treating propagation delay as invisible.
Overwriting rules without an audit or rollback model.
Ignoring how SDKs recover after missing distribution events.

Reveal stretch goalsOptional ways to extend the deliverable

Design staged percentage rollout with deterministic assignment.
Add rule simulation against recorded, privacy-safe attributes.

Optional solution worksheet

Prepare your approach for discussion.

Capture your reasoning, then copy a plain-text summary you control.

Session only - not saved.Refreshing or leaving this page clears every worksheet field and self-review choice. Nothing is uploaded or officially submitted.

Approach summary

Key decision

Important tradeoff

Failure mode or edge case

Validation or testing plan

Reflection

Optional solution URLThis link remains in the current browser session and is included only if you copy the summary.

Nothing is sent when you copy.

Optional community discussion

Compare reasoning, not rankings.

Discuss this challenge in the Engineering Foundry community if you choose. Share your own work, explain assumptions, critique ideas rather than people, and remove confidential or proprietary material.

Discuss this challenge

Loading Engineering Foundry…

System Design · Engineering Foundry challenge level: Advanced

Multi-Region Feature Flag Control Plane

Design a control plane and read path that distribute flag changes globally without making application requests depend on a central service.

All challengesSuggested practice time: 80 minutes

Original prompt

What you are building

Engineering Foundry originalNo company association

Deliverables

Assumptions and service-level objectives.
Control-plane and data-plane architecture.
Rule/version data model and distribution events.
APIs or configuration interfaces.
Reliability, rollback, and observability strategy.
Key tradeoffs and unresolved risks.

Constraints

Application evaluation must not require a network call on every request.
Rules must be auditable and tenant-isolated.
A bad rule must be recoverable without editing it in place.
Regional disconnection must have an explicit behavior.

Success criteria

Control and evaluation paths are separated.
Versioning and rollback behavior are concrete.
Propagation and stale-read risks are observable.
The design states how it behaves during regional and distribution failures.

Suggested workflow

Define ordinary and emergency change objectives.
Separate operator writes from application evaluation.
Model immutable versions and activation state.
Trace rollout, rollback, and regional partition scenarios.
Add observability and operational controls.

Attempt before reveal

Build your deliverable independently.

Use the prompt, constraints, and workflow first. Guidance offers considerations and questions - not a canonical solution.

Reveal guidanceDirections and questions to test

One direction to explore

Treat published rule versions as immutable artifacts and activation as a separate decision.
Compare streaming distribution with periodic snapshot recovery.
Ask which component can safely initiate an emergency disable when the primary control region is unavailable.

Reveal common mistakesFailure patterns worth checking

Putting a synchronous central lookup on the request path.
Treating propagation delay as invisible.
Overwriting rules without an audit or rollback model.
Ignoring how SDKs recover after missing distribution events.

Reveal stretch goalsOptional ways to extend the deliverable

Design staged percentage rollout with deterministic assignment.
Add rule simulation against recorded, privacy-safe attributes.

Optional solution worksheet

Prepare your approach for discussion.

Capture your reasoning, then copy a plain-text summary you control.

Session only - not saved.Refreshing or leaving this page clears every worksheet field and self-review choice. Nothing is uploaded or officially submitted.

Approach summary

Key decision

Important tradeoff

Failure mode or edge case

Validation or testing plan

Reflection

Optional solution URLThis link remains in the current browser session and is included only if you copy the summary.

Nothing is sent when you copy.

Optional community discussion

Compare reasoning, not rankings.

Discuss this challenge

Multi-Region Feature Flag Control Plane

What you are building

Deliverables

Constraints

Success criteria

Suggested workflow

Build your deliverable independently.

One direction to explore

System Design deliverable self-review

Prepare your approach for discussion.

Compare reasoning, not rankings.

Multi-Region Feature Flag Control Plane

What you are building

Deliverables

Constraints

Success criteria

Suggested workflow

Build your deliverable independently.

One direction to explore

System Design deliverable self-review

Prepare your approach for discussion.

Compare reasoning, not rankings.