[Improve][SDK] Fix non-atomic worker selection in getWorker in the DataProxy Go SDK

### Description

In the Dataproxy Go SDK, concurrent Send calls can start the round-robin scan from the same worker index, so the load is not evenly distributed across workers under high concurrency.

### InLong Component

InLong SDK

### Are you willing to submit PR?

- [x] Yes, I am willing to submit a PR!

### Code of Conduct

- [x] I agree to follow this project's [Code of Conduct](https://www.apache.org/foundation/policies/conduct)

Locate getWorker in the Dataproxy Go SDK and trace how concurrent Send calls choose a worker. Reproduce the behavior with concurrent sends, then verify that worker selection is atomic and distributes calls evenly under high concurrency.

Written by the indexing model from the issue text.