🗓️ 04032025 2318
📎
flink_checkpointing
- Ensures fault tolerance by saving operator flink_state periodically
- If a failure occurs, Flink restores state from the last successful checkpoint.
- Works with stateful functions that retain data across events.
Prerequisites for Checkpointing
| prerequisite | usage | examples |
|---|---|---|
| Durable data sources | For replaying records | Kafka, RabbitMQ, Kinesis, etc. |
| State storage | saving checkpoint snapshots | HDFS, S3, Ceph, etc. |