Java Spring Boot – SSE-II Offline Drive

Question Set
(Backend + Database + System Design + DSA)

Backend-Focused Questions (Spring Boot

+ Java Ecosystem)
1.​ Explain Filters, Interceptors, and AOP Advice in Spring Boot.​
Give real-world use cases and clearly differentiate where each fits in the request
lifecycle.​

2.​ When exactly do you choose Filter vs HandlerInterceptor vs AOP?​

Draw the exact execution pipeline for a Spring MVC request.​

3.​ Explain SOLID principles with Java-centric examples (Spring Beans, Services,
Repositories).​

4.​ Walk through how dependency injection works in Spring.​

How does the container resolve beans? Explain bean scopes (singleton,
prototype, request/session scope).​

5.​ Repository Pattern vs JPA Query Methods vs JPQL vs Native SQL — when do you
choose which?​

6.​ Explain optimistic locking vs pessimistic locking in Hibernate.​

Which one is better for high-contention write-heavy systems?​

7.​ How do you implement scheduled jobs and distributed background workers in Spring?​
(ScheduledExecutor, Spring Scheduler, Quartz, Kafka consumers)​

8.​ API versioning in Spring Boot — (URI, header-based, content-negotiation). Which scales
best long-term?​
9.​ Explain distributed caching with Spring Cache + Redis.​
How do you design cache invalidation for a multi-instance deployment?​

10.​What is eventual consistency? Give a Spring/Kafka-based scenario where it is preferred

over strict consistency.​

11.​Explain the Saga Pattern and how it would be implemented in Java using Kafka/Outbox
pattern.​

12.​Modular monolith vs microservices in Spring.​

How do module boundaries change? What about shared DTOs vs independent
bounded contexts?​

13.​What is an idempotent API?​

Show how to implement an idempotent POST using database constraints or Redis
locks.​

14.​Explain complete lifecycle of a Spring Boot request:​

Filter → Interceptor → Controller → Service → Repository → ExceptionHandler.​

15.​Explain how Hibernate’s MVCC works internally.​

How does it enable high concurrency?​

16.​What are ACID properties?​

How does PostgreSQL ensure them under concurrent writes with Spring Data JPA in
play?​

17.​How do you analyze slow queries in Java/Hibernate?​

How do you interpret EXPLAIN ANALYZE and Hibernate logs?​

18.​When do you prefer partial indexes, covering indexes, composite indexes?​

19.​Explain transaction behavior under READ COMMITTED vs REPEATABLE READ with

Spring @Transactional.​

20.​Design an audit logging system using JPA events + database triggers.​

What are the pros/cons of each?​

21.​SSE vs WebSockets vs Long Polling — exact differences in Java (Spring WebFlux or

Servlet stack).​

22.​Explain how Redis Pub/Sub or Kafka helps scale WebSocket events across multiple
pods.​
23.​How do you authenticate WebSocket connections securely in Spring?​

24.​What problems arise when broadcasting large payloads to thousands of clients?​

How to mitigate memory pressure?​

25.​Implement role-based + permission-based authorization using Spring Security.​

How do you create custom annotations like @HasPermission()?​

26.​Design a global exception handler using @ControllerAdvice for consistent API

responses.​

27.​Explain common Java backend security vulnerabilities (XXE, SSRF, JWT misuse,
deserialization attacks).​

28.​How do you store API secrets securely in Java?​

(Vault, AWS SM, encrypted environment configs)​

29.​Explain refresh token rotation and preventing replay attacks in Spring Security.​

30.​Design a resilient API integration layer using Resilience4j (retry, circuit breaker, fallback).​

31.​Redis cache vs in-memory cache (Caffeine) in multi-node Spring Boot clusters.​

Explain correctness issues.​

32.​How do you design a job queue with retries, DLQ, delay scheduling in Java?​
(Kafka, RabbitMQ, custom scheduler)​

33.​How do you design cache invalidation for create/update/delete flows?​

34.​How do you prevent thread starvation when a downstream API is slow or rate-limited?​

35.​Explain how Kubernetes HPA scales Java workloads using CPU/memory/custom

metrics.​

36.​During rolling deployments, what problems arise with schema mismatches?​

How do you prevent them?​

37.​How do you design a blue-green or canary deployment pipeline for Java apps?​

38.​Importance of liveness/readiness/startup probes for Spring Boot containers.​

39.​How do you generate typed API clients in Java + React?​

(OpenAPI/Swagger Codegen)​
 40.​How do React Suspense / Server Components reduce backend load?​
Explain from a Java API design perspective.​

41.​How do you optimize large list endpoints?​

Pagination, cursor-based pagination, projections, DTO slicing.​

42.​How to avoid over-fetching/under-fetching in complex React + Spring APIs?​

(GraphQL vs REST vs selective DTOs)​

Database Design Use-Cases 1. OAuth2 / SSO Identity

Service

● Ask the candidate to design a multi-provider SSO system supporting Google, GitHub, etc.
They must model user identities, sessions, and token lifecycles.​
● Tables Expected: users, providers, identities, refresh tokens, sessions.​
● Query to ask after DB design: List all active Google sessions for users who signed in last
14 days.​
● Sample Answer:

SELECT s.*​
FROM sessions s​
JOIN identities i ON s.identity_id = [Link]​
JOIN providers p ON i.provider_id = [Link]​
JOIN users u ON i.user_id = [Link]​
WHERE [Link] = 'google'​
AND s.is_active = TRUE​
AND u.last_login_at >= NOW() - INTERVAL '14 days';

2. Short Link Platform with Analytics

● Candidate should design a Bitly-style shortener with click tracking, geo metadata, and device
analytics. Schema should consider high-read, high-write workloads.​
● Tables Expected: links, redirects, geo events, device meta.​
● Query to ask after DB design: Top 5 countries for a link in the past 7 days.​
● Sample Answer:

SELECT [Link], COUNT(*) AS clicks​

FROM geo_events ge​
 WHERE ge.link_id = $1​
AND ge.created_at >= NOW() - INTERVAL '7 days'​
GROUP BY [Link]​
ORDER BY clicks DESC​
LIMIT 5;

3. Feature Flags / Experimentation

● Design a scalable experimentation system that assigns users to variants, stores rules, and
supports gradual rollouts.​
● Tables Expected: flags, variants, rules, assignments.​
● Query: Variant distribution for each active flag.​
● Sample Answer:

SELECT [Link] AS flag_name,​

[Link] AS variant_name,​
COUNT(a.user_id) AS user_count​
FROM flags f​
JOIN variants v ON v.flag_id = [Link]​
LEFT JOIN assignments a ON a.variant_id = [Link]​
WHERE f.is_active = TRUE​
GROUP BY [Link], [Link]​
ORDER BY [Link], [Link];

4. SaaS Billing (Subscriptions + Usage Metering)

● Design tables for tenants, subscription plans, usage logs, billing cycles, invoice generation,
and payments.​
● Tables Expected: tenants, subscriptions, usage, invoices.​
● Query: MRR by tenant for the current month.​
● Sample Answer:

SELECT [Link] AS tenant_id,​

SUM([Link]) AS mrr​
FROM tenants t​
JOIN subscriptions s ON s.tenant_id = [Link]​
JOIN invoices i ON i.subscription_id = [Link]​
WHERE DATE_TRUNC('month', i.created_at) = DATE_TRUNC('month', NOW())​
 GROUP BY [Link];

5. Retryable Distributed Job Queue

● Model a resilient job processing system with retries, delayed jobs, attempt logs, and a
dead-letter queue.​
● Tables Expected: jobs, attempts, schedules, dlq.​
● Query: Jobs that failed >3 times in the last 24 hours.​
● Sample Answer:

SELECT [Link],​
COUNT([Link]) AS failed_attempts​
FROM jobs j​
JOIN attempts a ON a.job_id = [Link]​
WHERE [Link] = 'failed'​
AND a.created_at >= NOW() - INTERVAL '24 hours'​
GROUP BY [Link]​
HAVING COUNT([Link]) > 3;

6. Notification Delivery System (Multi-channel: Email, SMS, Push)

● Design a system supporting various notification channels, user preferences, delivery tracking,
and retry logic.​
● Tables Expected: users, notifications, channels, delivery_attempts, provider_responses.​
● Query: Users who failed to receive >3 SMS notifications in last 48 hours.​
● Sample Answer:

SELECT [Link] AS user_id​

FROM users u​
JOIN notifications n ON n.user_id = [Link]​
JOIN channels c ON n.channel_id = [Link]​
JOIN delivery_attempts da ON da.notification_id = [Link]​
WHERE [Link] = 'sms'​
AND [Link] = 'failed'​
AND da.created_at >= NOW() - INTERVAL '48 hours'​
GROUP BY [Link]​
HAVING COUNT([Link]) > 3;
DSA Questions (Java solutions)
1.​ Longest substring without repeating characters (sliding window).​

2.​ Minimum size subarray sum ≥ target (two pointers).​

3.​ Rotate array right by K (in-place reversal).​

4.​ Check if one string is a rotation of another.​

5.​ Optimized anagram check (character array count).​

6.​ Merge two sorted arrays without extra space.​

7.​ Find all anagrams of a pattern in a string (sliding window).​

8.​ Longest palindromic substring (expand around center / DP).​

Java Streams & Collectors Questions

1. Grouping + Aggregation (Real Backend Scenario)

Given a list of Transaction objects, write a Streams pipeline to:

●​ filter only SUCCESS transactions​

●​ group by accountNo​

●​ compute total credit and total debit separately​

●​ return a Map<String, AccountSummary>​

Explain pitfalls: null fields, negative values, rounding, BigDecimal vs double, concurrency.
2. map() vs flatMap() — Practical Backend Example

Explain the difference between map() and flatMap()​

with real examples such as:

●​ flattening nested DTOs ([Link]().stream())​

●​ mapping entities to response DTOs​

●​ flattening tags from multiple products​

3. Handling Duplicate Keys in toMap()

What happens when [Link]() receives duplicate keys?​

Rewrite with:

●​ merge function​

●​ and choosing between HashMap, LinkedHashMap, and TreeMap.​

4. Parallel Streams — When They Are Worse

Explain why parallel streams can be slower than sequential streams:

●​ ForkJoinPool overhead​

●​ shared mutable state​

●​ IO-bound operations​

●​ false sharing​

●​ ORDERED streams becoming bottlenecks​

Give a real backend example where parallel streams are harmful.​

5. Employee Highest Salary Per Department (Max + Grouping)

Write a Streams pipeline to group employees by department and return​
the highest-salary employee per department.​
Avoid unsafe [Link]() and return a clean DTO.

6. Custom Collector Using [Link]()

Implement a custom collector to compute average salary:

●​ accumulator​

●​ combiner​

●​ finisher​
Explain why you may prefer this over [Link]() in
performance-critical code.​

7. List of High Earners (> Average Salary) Using Streams

Using Streams only:

●​ compute org-wide average salary (efficiently, without multiple passes)​

●​ return employees whose salary > average​

Explain pitfalls: overflow, numeric precision, repeated iteration.