Arcitura Certified Microservice Architect (S90.MSA) — Questions and Answers

Question 1: A team is deploying microservices in a container orchestration environment like Kubernetes. When a client service needs to communicate with a provider service, it sends a request to a stable virtual IP address managed by the platform. The platform then routes this request to a healthy instance of the provider service. Which service discovery pattern does this scenario describe?

Self-Registration
Server-Side Discovery (Correct answer)
Client-Side Discovery
Third-Party Registration

Correct answer: Server-Side Discovery

This scenario describes Server-Side Discovery. The client sends a request to a logical endpoint (the Kubernetes Service's stable IP), and the platform's networking layer (acting as a server-side proxy/load balancer) is responsible for discovering a healthy service instance and forwarding the request. The client is completely abstracted from the complexity of looking up individual service instances.

Question 2: What is a key drawback of the Database-per-Service pattern?

It forces a shared schema
Implementing queries and transactions spanning services is harder (Correct answer)
It prevents independent deployment
It eliminates data ownership

Correct answer: Implementing queries and transactions spanning services is harder

With separate databases, cross-service queries and consistency require patterns like saga or composition.

Question 3: When a new pod replaces a crashed one in Kubernetes, clients can still reach the workload because:

DNS is disabled
They connect via the Service's stable name/IP, not the pod IP (Correct answer)
They restart automatically
The pod keeps the same IP

Correct answer: They connect via the Service's stable name/IP, not the pod IP

Clients target the stable Service abstraction, which is updated to point at the new pod.

Question 4: Why is centralized logging important in microservices?

It replaces the need for testing
It reduces the number of services
Requests span many services, so aggregated logs are needed to trace behavior (Correct answer)
Logs are useless in distributed systems

Correct answer: Requests span many services, so aggregated logs are needed to trace behavior

Aggregated, correlated logs are essential because a request flows through many services.

Question 5: Which pattern is commonly used to maintain data consistency across services without distributed transactions?

Saga pattern (Correct answer)
Two-phase commit lock
Global table lock
Synchronous join query

Correct answer: Saga pattern

The Saga pattern coordinates a sequence of local transactions with compensating actions instead of a distributed transaction.

Question 6: Which pattern maintains a denormalized read model updated from events?

Bulkhead
Strangler Fig
Ambassador
Materialized View (CQRS read model) (Correct answer)

Correct answer: Materialized View (CQRS read model)

A materialized view precomputes a query-optimized read model kept current via events.

Question 7: When the Circuit Breaker is in the 'half-open' state, what happens?

A limited number of trial requests test if the dependency recovered (Correct answer)
The service restarts
All requests are blocked permanently
All requests pass freely

Correct answer: A limited number of trial requests test if the dependency recovered

Half-open lets a few probe requests through to decide whether to close or re-open the breaker.

Question 8: Which pattern addresses reliably publishing events as part of a database transaction?

Transactional Outbox (Correct answer)
Sidecar
Bulkhead
Circuit Breaker

Correct answer: Transactional Outbox

The Transactional Outbox writes events to a table in the same transaction, then relays them, avoiding dual-write inconsistency.

Question 9: What does a multi-stage Docker build primarily help achieve?

Built-in service discovery
Smaller final image size (Correct answer)
Faster container startup
Automatic horizontal scaling

Correct answer: Smaller final image size

Multi-stage builds discard build-time dependencies, producing a leaner final image.

Question 10: Which of these is the main reason microservices avoid distributed ACID transactions across services?

They are faster than local transactions
They reduce scalability and increase coupling between services (Correct answer)
They are required by REST APIs
They guarantee zero latency

Correct answer: They reduce scalability and increase coupling between services

Distributed transactions like 2PC lock resources across services, harming scalability and availability.

Question 11: Which principle best describes how each microservice should be structured around a specific business capability?

Shared global state
Domain-driven bounded contexts (Correct answer)
Database centralization
Monolithic layering

Correct answer: Domain-driven bounded contexts

Microservices are organized around bounded contexts from domain-driven design, each owning a single business capability.

Question 12: Which of the following is a primary responsibility of a container orchestration platform like Kubernetes?

Managing the application's database schema and migrations.
Building container images from a source code repository.
Automating the deployment, scaling, and health management of containers across a cluster. (Correct answer)
Writing the application's business logic and API endpoints.

Correct answer: Automating the deployment, scaling, and health management of containers across a cluster.

A container orchestrator's main purpose is to manage the lifecycle of containers at scale. This includes scheduling containers onto available nodes, automatically scaling services up or down based on demand, restarting failed containers (self-healing), and handling service discovery and load balancing.

Question 13: Which pattern provides a single entry point that routes client requests to backend services?

CQRS
Event Sourcing
Circuit Breaker
API Gateway (Correct answer)

Correct answer: API Gateway

An API Gateway aggregates and routes requests, handling cross-cutting concerns.

Question 14: In Kubernetes, what is the purpose of a liveness probe?

To check if a container should receive traffic
To scale Pods based on CPU
To restart a container that has become unhealthy (Correct answer)
To mount persistent storage

Correct answer: To restart a container that has become unhealthy

A liveness probe detects a stuck container and triggers a restart.

Question 15: What is a compensating transaction in a saga?

A retry of a failed read
An operation that undoes a previously completed step (Correct answer)
A load-balancing decision
A cache invalidation

Correct answer: An operation that undoes a previously completed step

Compensating transactions reverse the effects of earlier steps when a saga must roll back.

Question 16: Which header is commonly injected by a gateway to enable distributed tracing across services?

User-Agent
A correlation/trace ID header (Correct answer)
Host
Content-Length

Correct answer: A correlation/trace ID header

Gateways add a correlation or trace ID so a request can be followed across multiple microservices in tracing tools.

Question 17: What is a typical trade-off of using an AP (availability-favoring) registry?

It cannot scale
It always returns errors
It may serve stale instance data during partitions (Correct answer)
It blocks all writes

Correct answer: It may serve stale instance data during partitions

Choosing availability means clients may occasionally receive outdated registration information.

Question 18: In the context of containerization, what is the primary purpose of a container image?

To provide a live, running, and isolated environment for an application.
To define the networking rules and load balancing strategy for inter-service communication.
To serve as a static, immutable package containing an application's code, runtime, and all dependencies. (Correct answer)
To monitor the real-time performance and resource consumption of a running microservice.

Correct answer: To serve as a static, immutable package containing an application's code, runtime, and all dependencies.

A container image is a read-only template or blueprint that packages up the application code along with all its necessary dependencies, libraries, and configuration files. A container is the runnable instance created from an image. The image itself is static and portable.

Question 19: In a microservices architecture, what does the 'Database per Service' pattern primarily ensure?

Data is never persisted to disk
Services bypass APIs to query each other's tables
Each service owns and encapsulates its own data store (Correct answer)
All services share a single relational database

Correct answer: Each service owns and encapsulates its own data store

Database per Service gives each microservice exclusive ownership of its data, preventing tight coupling at the data layer.

Question 20: An organization is migrating a legacy monolithic application to microservices. They need a way to synchronize data from the monolith's database to the new microservices' databases in near real-time without modifying the monolith's application code. Which pattern is best suited for this purpose?

Transactional Outbox
Change Data Capture (CDC) (Correct answer)
API Composition
Strangler Fig Pattern

Correct answer: Change Data Capture (CDC)

Change Data Capture (CDC) is a pattern used to monitor and capture row-level changes (inserts, updates, deletes) in a database's transaction logs and stream these changes as events to other systems. This allows new microservices to stay synchronized with the legacy database without requiring intrusive changes to the original application's code.

Question 21: How does a registry typically detect that an instance has died without deregistering?

Through health checks or heartbeat timeouts (Correct answer)
By restarting the instance
By scanning source code
It never detects it

Correct answer: Through health checks or heartbeat timeouts

Missed heartbeats or failed health checks cause the registry to expire stale entries.

Question 22: What does Kubernetes Horizontal Pod Autoscaler (HPA) adjust automatically?

CPU limits per container
The number of Pod replicas (Correct answer)
The size of persistent volumes
Node memory allocation

Correct answer: The number of Pod replicas

HPA scales the replica count up or down based on observed metrics like CPU.

Question 23: What does idempotency guarantee in service-to-service messaging?

Strong ordering across all services
Messages are never delivered
Zero network latency
Processing the same message multiple times has the same effect as once (Correct answer)

Correct answer: Processing the same message multiple times has the same effect as once

Idempotency ensures repeated delivery of the same message produces the same result, vital for at-least-once messaging.

Question 24: A development team is working on a legacy monolithic application and wants to gradually migrate its functionality to a new microservices architecture without performing a risky 'big bang' rewrite. They plan to incrementally build new microservices and route traffic to them, while the old monolith continues to handle the remaining functionality. Which pattern facilitates this gradual migration strategy?

Strangler Fig Pattern (Correct answer)
Sidecar Pattern
Circuit Breaker Pattern
API Composition Pattern

Correct answer: Strangler Fig Pattern

The Strangler Fig pattern is an architectural approach for incrementally migrating a legacy system. It involves creating a new application (the 'strangler') around the old one and gradually replacing pieces of the monolith's functionality with new microservices. An intermediary layer, often a proxy or API gateway, intercepts requests and routes them to either the new microservice or the old monolith, allowing for a safe, phased migration.

Question 25: Which pattern lets you incrementally migrate a monolith by intercepting and replacing functionality over time?

Sidecar
Bulkhead
Ambassador
Strangler Fig (Correct answer)

Correct answer: Strangler Fig

The Strangler Fig pattern gradually replaces legacy features behind a routing facade.

Question 26: Which of the following describes the primary role of an API Gateway in a microservices architecture?

To act as a single entry point for clients, handling routing, security, and other cross-cutting concerns. (Correct answer)
To orchestrate complex, multi-service business transactions using the Saga pattern.
To provide a persistent storage layer shared by all microservices.
To manage the deployment and scaling of individual microservice instances.

Correct answer: To act as a single entry point for clients, handling routing, security, and other cross-cutting concerns.

The fundamental purpose of an API Gateway is to serve as a reverse proxy and unified entry point for all external client requests. It abstracts the underlying microservice architecture, routing requests to the appropriate services while handling concerns like authentication, rate limiting, and SSL termination. It decouples clients from the internal service structure. Service scaling, transaction orchestration, and data storage are handled by other components in the ecosystem (e.g., container orchestrators, the services themselves).

Question 27: In service mesh terminology, what handles east-west (service-to-service) traffic while a gateway handles north-south traffic?

Sidecar proxies (Correct answer)
DNS servers
CDN edge nodes
Load balancers only

Correct answer: Sidecar proxies

Service meshes use sidecar proxies for internal east-west traffic, whereas the gateway manages external north-south traffic.

Question 28: When it comes to microservices, how does a backing service fit in?

When a microservice can't handle the compute demand anymore, it's shut down.
It acts as a dedicated service that provides essential functionality required by a microservice (Correct answer)
It synchronizes network activity among microservices.
It prevents the failure of a microservice.

Correct answer: It acts as a dedicated service that provides essential functionality required by a microservice

A backing service is any service that a microservice consumes over the network to perform its normal operations. These are typically external resources like databases, message queues, caching systems, or third-party APIs that provide essential infrastructure or data. They are managed independently and are crucial for the microservice's functionality.

Question 29: A microservice needs to handle cross-cutting concerns like collecting detailed logs and metrics without bloating the primary application's code. The team wants to deploy a separate, specialized container for these tasks alongside every instance of the main microservice, sharing the same network space and lifecycle. Which container design pattern is best suited for this scenario?

Strangler Fig Pattern
Ambassador Pattern
Blue/Green Deployment
Sidecar Pattern (Correct answer)

Correct answer: Sidecar Pattern

The Sidecar pattern involves co-locating a helper container with the main application container (often in the same Kubernetes Pod). This allows the sidecar to augment the main application by handling tasks like logging, monitoring, or acting as a proxy, without being tightly coupled to the application's code.

Question 30: What is the purpose of request/response transformation at the gateway?

Adapting payload formats or protocols between clients and services (Correct answer)
Compiling source code
Storing user passwords
Provisioning servers

Correct answer: Adapting payload formats or protocols between clients and services

Transformation lets the gateway reshape payloads or translate protocols (e.g., REST to gRPC) so clients and services need not match exactly.

Arcitura Certified Microservice Architect (S90.MSA)

The S90.MSA certification validates expertise in microservice architecture design patterns, containerization, data management, CI/CD deployment strategies, and API gateway management across cloud-native environments.

Questions

35 min

Time Limit

80%

To Pass

Checking for saved progress...

Exam Rules

You can skip questions and return to them later
Flag questions for review before submitting
No feedback shown until you submit the entire exam
Unanswered questions count as wrong — answer everything
10 pretest questions are mixed in and don't affect your score
Timer auto-submits when time runs out
Your progress is auto-saved every 30 seconds