fix(sqs-batch-manager): use original queueUrl instead of batchKey in batch requests

[anuq]

Problem

Fixes #6478

When a SendMessageRequest, DeleteMessageRequest, or ChangeMessageVisibilityRequest has an overrideConfiguration set, getBatchKey() correctly returns queueUrl + overrideConfig.hashCode() as a composite key to group requests for batching. However, createXxxBatchRequest() was using this same composite string as the QueueUrl in the actual SQS API call, producing a non-existent URL like:

https://sqs.us-west-2.amazonaws.com/123/my-queue-461286369

This causes QueueDoesNotExistException for any caller that sets an overrideConfiguration on their request — notably the SQS extended client library which appends a user-agent string via overrideConfiguration.

Fix

Extract queueUrl from identifiedRequests.get(0).message().queueUrl() (safe because all requests in a batch share the same queueUrl by the grouping invariant) and use that instead of batchKey for the QueueUrl field in the batch API call.

Affected classes:

• SendMessageBatchManager
• DeleteMessageBatchManager
• ChangeMessageVisibilityBatchManager

Testing

Added regression tests to BaseSqsBatchManagerTest for all three operations that verify the WireMock-captured request body contains the original QueueUrl (not the hash-suffixed batchKey) when overrideConfiguration is set.

Checklist

• Includes tests to exercise the new behavior
• My code follows the code style of this project
• Git commit message is detailed and includes context behind the change
• Issue number is referenced (#6478)
• I confirm that this pull request can be released under the Apache 2 license
• Local run of ./mvnw package succeeds