News, guides, and engineering deep dives.
Practical guidance on Kafka, Flink, Iceberg, and real-time data.
Kafka in Docker
Running Kafka in Docker: the official images, reliable Docker Compose topologies, the advertised.listeners trap that breaks local connections, and why a single-node container is not a deployment.
Kafka offsets
A Kafka offset is a record's position in its partition, and a committed offset is a consumer group's bookmark. Reading lag from CURRENT-OFFSET and LOG-END-OFFSET, plus the reset strategies.
Kafka producers in production
A Kafka producer appends records to topic partitions. Configuration and tuning with real numbers, idempotence and delivery guarantees, and the client-library decision that quietly matters most.
Kafka Streams
Kafka Streams is a Java library for stream processing that runs inside your application, with state in local RocksDB stores and changelog topics. The operational realities, and where Flink wins.
Kafka Streams documentation, mapped
The Kafka Streams documentation divides into four layers: API reference, configuration surface, state store internals, and the upgrade guide. Which layer answers which production question.
Kafka topics
A Kafka topic is a named, append-only log, and in production it is the unit you operate on. The CLI verbs, partition and replication mechanics, retention, compaction and the troubleshooting moves.
A Kafka topic example, fully specified
A production Kafka topic example: the exact creation command with durability properties, the same topic as Terraform and Strimzi code, the naming convention that scales, and the schema contract.
Kafka topic vs partition
A topic is the logical name; a partition is the physical log. How replication, ordering, parallelism and key hashing follow the physical unit, and why keyed topics lock their count.
A production Kafka tutorial
A Kafka tutorial for people who run it in production: zero-downtime upgrades, broker tuning, layered security, troubleshooting signals, and the client settings that decide delivery guarantees.