Seeking Advice on Minimizing Downtime During Application Logic Upgrades in Flink Cluster

[Ruibin Xing <xi...@gmail.com>]

Hi Community,

I need some advice on upgrading the application logic in our Apache Flink
cluster without causing any downtime. We're using Flink for our most
important data pipeline, and we need to keep upgrading the application
logic to meet our business needs. However, we're worried about the downtime
that could affect our users.

I know it might not be possible to achieve zero downtime in the current
Flink version, but I think many users have faced the same problem. I'm
wondering if it's possible to achieve that in future, and if so, what's the
possible directions?

Thanks a lot for your help!

Best regards,
Ruibin