loaded and allows to determine which state actually had the run-away and what
might have caused duplication
I hope this helps
Thias
From: Frederic Leger
Sent: Monday, August 28, 2023 12:30 PM
To: user@flink.apache.org
Subject: Checkpoint/savepoint _metadata
⚠EXTERNAL MESSAGE – CAUTION: Think
Hi team,
We use flink 1.16.0 with openjdk-11-jre mainly to run streaming jobs.
We do checkpoints with 2 min interval and savepoint when deploying new job
version.
We also use rocksdb state backend for most of them.
We had a streaming job running for long without any issue and during a new
deploym
