🗓️ 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​

prerequisiteusageexamples
Durable data sourcesFor replaying recordsKafka, RabbitMQ, Kinesis, etc.
State storagesaving checkpoint snapshotsHDFS, S3, Ceph, etc.

References​