Architecture Pattern Match Game

Know the cards before entering the arena.

This trainer is about reading dominant system pressure, not memorizing pattern names. Pick your mode, study the cards, then enter a focused seven-round run.

See the pattern lobbyLinkedIn Share on X

7 scenario rounds

Practice mode

Architecture Pattern Match Game hero artwork

Pattern lobby

Pick your mode, then enter the arena.

Pick your help level

Pattern card

Strangler Fig

Migration

Put a new facade around the old building, then replace rooms one at a time.

Incremental legacy modernization

Avoid when the old system is simple enough to replace directly and the migration window is low risk.

Pattern card

Saga

Rollback

A travel itinerary with cancellation rules for each booking leg.

Distributed workflows across several services

Avoid when you need strict immediate consistency across every step and cannot tolerate intermediate states.

Pattern card

CQRS

Read shape

One kitchen takes orders, another line plates dishes for speed.

Heavy read traffic with richer query needs than the write model supports

Avoid when CRUD is sufficient and read/write scaling pressure is still modest.

Pattern card

Event-Driven

Fan-out

A radio broadcast that many listeners can tune into at the same time.

Fan-out workflows with many downstream consumers

Avoid when the caller needs a synchronous yes-or-no answer right now.

Pattern card

Sidecar

Operational placement

A support van driving next to the main race car handling fuel, telemetry, and radios.

Cross-cutting operational concerns such as logging, mTLS, or proxying

Avoid when the concern is deeply application-specific or the platform cost outweighs the reuse benefit.

Pattern card

Circuit Breaker

Resilience

Trip the fuse before the whole house overheats.

Protecting against cascading dependency failure

Avoid when the dependency is not actually optional and no safe degraded mode exists.

Pattern card

Cache-Aside

Hot reads

Keep the most-used books on the desk and fetch from the archive only when needed.

Read-heavy workloads with repeat access patterns

Avoid when data freshness requirements are stricter than the cache lifecycle can support.