Skip to content
#

savepoint

Here are 10 public repositories matching this topic...

Where a Kafka to Iceberg pipeline stops being exactly-once: Flink 1.20, Iceberg 1.10, Kafka 4.3 and Trino 478 in Docker Compose. Turning off exactly-once checkpointing changed nothing; an old savepoint froze the table for about a minute at zero consumer lag, then committed 12,000 duplicates; checkpoint interval sets file size, not target file size.

  • Updated Aug 30, 2026
  • Python

Add this topic to your repo

To associate your repository with the savepoint topic, visit your repo's landing page and select "manage topics."

Learn more