[Bug]: GBK operation fails on OutOfMemory with SparkRunner

[JozoVilcek]

What happened?

When upgrading from 2.33.0 to 2.61.0, job on spark runner start to fail on OutOfMemory.
The culprit is that this change: #15637 which effectively disabled memory sensitive GBK translation using secondary sort introduced by: https://issues.apache.org/jira/browse/BEAM-5392

I propose to restore the memory sensitive translation by considering it first and only if it can not be used consider default GBK translation and it's global window optimisation.

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 YAML
• Component: Beam examples
• Component: Beam playground
• Component: Beam katas
• Component: Website
• Component: Infrastructure
• Component: Spark Runner
• Component: Flink Runner
• Component: Samza Runner
• Component: Twister2 Runner
• Component: Hazelcast Jet Runner
• Component: Google Cloud Dataflow Runner