What happened?

Hi,

We updated a Pubsub streaming job on Dataflow from 2.46.0 to 2.49.0. See these memory diagrams:

2.46.0 memory utilisation:

2.49.0 memory utilisation:

We sent back on the 2.46.0 for this job as workers were running out of memory and a lot of lag was introduced.

Do you have an explanation? What changed on memory management between 2.46.0 and 2.49.0?

Issue Priority

Priority: 2 (default / most bugs should be filed as P2)

Issue Components

• Component: Python SDK
• Component: Java SDK
• Component: Go SDK
• Component: Typescript SDK
• Component: IO connector
• Component: Beam examples
• Component: Beam playground
• Component: Beam katas
• Component: Website
• Component: Spark Runner
• Component: Flink Runner
• Component: Samza Runner
• Component: Twister2 Runner
• Component: Hazelcast Jet Runner
• Component: Google Cloud Dataflow Runner