Spring Boot Interview Questions

You probably use Spring Boot every day without reading its startup log, and interviewers know it, which is why they open here. This section covers the questions that check whether you understand the framework or only its defaults.

• What does @SpringBootApplication actually combine, and what would break if you removed one part?
• How does autoconfiguration decide what to configure, and what role do conditional annotations play?
• What is a starter, and what is inside one besides dependencies?
• How do you override or disable a specific autoconfiguration?
• Where does an embedded server come from, and how would you swap it?

A good answer connects the mechanism to something you did: a bean you had to exclude, a starter you replaced, a startup failure you traced through the condition evaluation report.

• Dependency injection. Constructor versus field injection, why constructor injection is preferred, and how to resolve two candidate beans of the same type.
• Bean lifecycle and scopes. Why singleton is the default, when request or prototype scope is correct, and what makes a singleton bean unsafe to share.
• Configuration. Property precedence, profiles per environment, and @ConfigurationProperties versus scattered value injection.
• Spring Data JPA. Derived query methods, the N plus 1 problem, lazy versus eager fetching, and when to drop to a native query.
• Transactions. What @Transactional proxies actually intercept, why a self-invocation silently skips the transaction, and how rollback rules work with checked exceptions.

The transaction proxy question is the single most common place a confident answer falls apart, so be ready to explain the proxy boundary rather than just naming the annotation.

Beyond the container, interviewers want the parts that show up in production. On the web layer: status code choices, validation, and centralized error handling with @ControllerAdvice instead of try-catch in every controller. On testing: the difference between a full @SpringBootTest and a slice such as a web-layer test, and why starting the whole context for one controller test is a slow habit. On operations: what Actuator exposes, which endpoints must not be public, and how health checks feed a deployment.

Consider a backend engineer interviewing for a mid-level role at a logistics company. She is asked why a scheduled job sometimes wrote partial data. Rather than guessing, she walks through the transaction boundary, notes the job called a transactional method on itself, and explains the proxy behavior that skipped it. The interviewer moved straight to an offer conversation, because the answer showed a debugging habit rather than a memorized definition.

Question sets for other languages and frameworks are collected under question banks.

Read your own service's startup log once and be able to narrate it. Then rehearse out loud, because the gap between knowing a concept and explaining it in forty seconds is where these rounds are decided. An AI interviewer on the /mock-interview page can ask the follow-ups you would not ask yourself.

During a spoken video interview, SubcueAI transcribes the interviewer's question and surfaces a suggested structure on your side, in the desktop overlay on macOS and Windows or in the Chromium browser extension side panel. No meeting bot joins the call, and nothing is injected into the meeting page. The honest limit is the coding round: if it moves onto a screen-monitored or proctored platform, live help is out of scope, and if you share your screen, everything on it is visible.