Apache Flink Streaming
Build stateful stream-processing jobs with Apache Flink. Writes DataStream and SQL jobs, manages windows, watermarks, and state, configures checkpointing, and tunes throughput for real-time pipelines.
This skill creates and operates Apache Flink streaming jobs. It writes DataStream API and Flink SQL pipelines, configures event-time windows, watermarks, and keyed state, sets up checkpointing and exactly-once guarantees, and tunes parallelism and throughput for low-latency real-time processing.
flink stream-processing real-time stateful data-engineering
When to use
Use when building Flink streaming jobs, configuring windows, watermarks, and state, setting up checkpointing, or tuning a real-time pipeline's throughput.
Examples
Windowed aggregation
Aggregate a stream
Write a Flink job that consumes a Kafka clickstream and computes 5-minute tumbling-window counts per page with event-time watermarks
Stateful processing
Use keyed state
Build a Flink DataStream job that tracks running session state per user and emits a session-summary event on timeout
Checkpointing setup
Add fault tolerance
Configure exactly-once checkpointing for my Flink job with a 30-second interval and RocksDB state backend