Java & Spring Developers India

We have 6 services and integration bugs keep slipping through. Has anyone used either in production? Which was easier for the team to adopt?

Two backend roles (SDE-2 and Senior), Bengaluru hybrid. Spring Boot, Kafka, Postgres, lots of interesting payments problems. Happy to answer questions here or in DMs.

Layer-based (controller/service/repository) or feature-based (invoice/, payment/, customer/)? Our codebase is layer-based and finding things is getting hard at ~200 classes.

We used H2 for integration tests for years. Every few months a query that passed on H2 failed on Postgres in prod. Moved to Testcontainers with the real Postgres image + reusable containers. Test suite went from 6 to 8 minutes, and we haven't had a dialect bug since. Tip: enable container reuse locally

For DTOs we've moved to records. Entities still use Lombok @Getter/@Setter. Is anyone going Lombok-free entirely? What did you replace @Builder with?

One runaway report query held a lock for 40 minutes and took down our payouts API last month. Now every service sets `options=-c statement_timeout=30s` on the Hikari datasource. Long-running jobs use a separate datasource with a higher limit. Should have done this years ago.

Curious where everyone is. We just finished moving everything to 21.

Not a checklist, but patterns I've seen in people who made the jump: 1. They ask 'what happens when this fails?' before anyone else does. 2. Their PRs are small and their descriptions explain why, not what. 3. They unblock others without being asked. 4. They can say 'I don't know, let me find out' comfortably.

Starting a new notifications service. Team knows Spring MVC well. Is there any reason to choose WebFlux now that we have virtual threads in Java 21?

Reviewed three architectures this month where a 6-person team runs 12 services. Every one of them had: • distributed transactions they didn't know they had • a shared DB anyway • a deploy pipeline more complex than the product A well-structured modular monolith (Spring Modulith is great for this) gets you 90% of the

Spent 2 hours on this today. My service method throws a custom checked exception after saving an entity, and the save is NOT rolled back. ```java @Transactional public void createInvoice(...) throws InvoiceException { repo.save(invoice);

We had duplicate payouts twice last year. Both times: client retried after a timeout, server had already processed the first call. What fixed it for good: 1. Every mutating API takes an Idempotency-Key header (client generates UUID). 2. We store (key, request_hash, response, status) in Postgres with a unique index on

Started this because most Java groups I'm in are either job spam or course ads. What this space is for: • Real questions from real projects (stack traces welcome) • Spring Boot / JPA / Kafka / JVM tuning discussions • Interview prep, but focused on understanding, not memorising