Add a typed client API to submit Issue Detection for traces

[!WARNING] Before submitting a PR, please make sure that:

• A maintainer has triaged this issue and applied the ready label
• This issue has no assignee
• No duplicate PR exists

PRs not meeting these requirements may be automatically closed.

Summary

MLflow has a UI-specific endpoint that starts issue detection for selected traces. It creates a run in the supplied experiment and returns job and run identifiers. Typed Python callers cannot invoke that behavior or retrieve the persisted issues.

Proposed change

Add a version-three Issue Detection submission RPC and typed client method. Reuse the existing server launch and job path. Return a typed submission handle with job_id and run_id. Expose typed experiment-scoped issue search by source_run_id.

Acceptance criteria

• The API preserves server-side credential handling.
• It accepts no raw provider key.
• It reuses the existing job and trace links.
• It preserves issues in the requested experiment.
• Server, transport, client, authorization, and proto tests pass.