Kafka transports history; Flink computes over streams#
Kafka → Flink → Database / Search / Another TopicFlink supports stateful stream processing, windows, event time, watermarks that represent event-time progress, and checkpoints for recoverable operator state. This bridge intentionally stops before APIs and topology internals.
| Concept | Purpose |
|---|---|
| Stateful operator | Remember data across events |
| Window | Aggregate a bounded time/key scope |
| Event time | Reason from when events occurred |
| Watermark | Estimate event-time progress amid late data |
| Checkpoint | Recover processing state and source positions |
Flink is Advanced for generic SWE, Important for infrastructure/streaming, and Must Know for many data-engineering roles. It is not required to understand Kafka.
Related practice problems
Use the manifest-backed prompts below to apply this lesson in a complete design.
- Stateful stream processingApache Flink Documentation
- Timely stream processingApache Flink Documentation
Reach this point to mark the lesson as read.