Skip to content

What broke in production,
why — and how I fixed it.

I build real-time Kafka and Flink platforms for core banking. This site is my engineering logbook: production failures traced to root cause, the architecture lessons behind them, and the resources I'd hand a new engineer on my team.

cluster: ggorantala.dev · 1 producer · 5 topics · 5 consumer groups

Why are you here? Subscribe to one.

producer.send() → topic log → consumer.poll()

KafkaProducer<Topic, Post>

Gopi Gorantala

15+ years of Java. Kafka, Flink & core-banking reconciliation in production.

acks=all · idempotence=true

ConsumerRecords<String, Post> records = consumer.poll(Duration.ofMillis(100)); // group.id=something-broke

live · auto.offset.reset=earliest

// directory

Everything on this site, in one table.

Section What's inside Come here when… Find it at
Error index Every production failure write-up, searchable by exception You have a stack trace open /blog/
Kafka errors Producer, consumer, broker and Kafka Streams failures Lag, timeouts, rejected records /tag/kafka-errors/
Flink errors Checkpoints, savepoints, resources, connectors A job won't start or keeps restarting /tag/flink-errors/
Core Java & JVM Class loading, serialization, concurrency, virtual threads The JVM itself is the problem /tag/java-errors/
Spring Boot Bean wiring, startup and configuration failures The context won't come up /tag/spring-boot/
Field notes Architecture decisions, trade-offs, life & career learnings You're designing, not debugging /blog/
Algorithms & DSA LeetCode walkthroughs, bit manipulation, data structures Interview prep or fundamentals /tag/dsa/
Books The reading list behind the work You want to go beyond code /productivity-books/
Courses (external site) Structured Java and Kafka tracks, Educative courses You want a learning path javahandbook.com ↗
About & newsletter Who I am, and one dissected failure in your inbox You want to follow along /about-me/ · /rss/

Hi, I'm Gopi.

Staff engineer, 15+ years in Java. I own streaming platforms end to end — Kafka topology, Flink state, CI/CD — and I still ship code. I also lead teams and mentor engineers.

The longer story →

One production failure, dissected.