Add new index on datasource and task_allocator_id for pending segments

[AmatyaAvadhanula]

Adds a new index on (datasource, task_allocator_id) for the druid_pendingSegments table to optimize fetching of pending segments by task_allocator_id for cleanup and transaction append / replace actions.

The index is added on columns that are (mostly) null, and is not expected to hinder the startup of the Overlord.
I have locally tested this out on a pending segments table with > 2M rows and this operation took 2.14s.

This PR has:

• been self-reviewed.
  • using the concurrency checklist (Remove this item if the PR doesn't have any relation to concurrency.)
• added documentation for new or modified features or behaviors.
• a release note entry in the PR description.
• added Javadocs for most classes and all non-trivial methods. Linked related entities via Javadoc links.
• added or updated version, license, or notice information in licenses.yaml
• added comments explaining the "why" and the intent of the code wherever would not be obvious for an unfamiliar reader.
• added unit tests or modified existing tests to cover new code paths, ensuring the threshold for code coverage is met.
• added integration tests.
• been tested in a test Druid cluster.

[kfaraz]

Thanks for the changes, @AmatyaAvadhanula !
Since the index is on datasource + taskAllocatorId, the method IndexerMetadataStorageCoordinator.deletePendingSegmentsForTaskAllocatorId() should accept datasource as an argument too.

[AmatyaAvadhanula]

Thanks for the review, @kfaraz