fix: throttle cleanup hang when broker uses default throttle configs

[bobelev]

Prevent throttle cleanup from hanging on brokers with dynamic default replication
throttle values by treating default/static values as cleared when expecting null,
allowing execution to advance.

Summary

1. Why: if cluster has DYNAMIC_DEFAULT_BROKER_CONFIG for replication.throttled settings cc fails to clear its own replication throttle.

Expected Behavior

1. Rebalancing works

Actual Behavior

1. ReplicationThrottleHelper.waitForConfigs gets stuck

Steps to Reproduce

1. set DYNAMIC_DEFAULT_BROKER_CONFIG leader.replication.throttled.rate cluster-wide
2. run rebalance
3. it'll get get stuck after a first batch of movements

Categorization

• documentation
• bugfix
• new feature
• refactor
• security/CVE
• other

[bobelev]

Please also backport this to migrate_to_kafka_3_0