Add async API support for ibm_db Integration

Summary

The IBM Db2 Haystack integration currently only supports synchronous
pipeline execution. This issue tracks adding full async support using
the AsyncConnection and AsyncCursor APIs introduced in ibm_db 3.3.0,
so the integration can participate in await pipeline.run_async(...) calls.

Motivation

Haystack's Pipeline now exposes run_async() / run_async_generator().
Components that implement async def run_async(...) can participate in
async pipeline execution. The IBM Db2 integration is currently sync-only,
which blocks users who want to run Db2-backed pipelines asynchronously
alongside other async components.

ibm_db 3.3.0 (released) introduces a native asyncio API
(AsyncConnection, AsyncCursor) that wraps synchronous IBM CLI calls
in asyncio.to_thread(), making it possible to add async support without
any additional runtime dependencies beyond Python's standard asyncio.

Start by reading the existing IBM Db2 integration's synchronous execution path and the ibm_db 3.3.0 AsyncConnection and AsyncCursor APIs. Trace how Haystack's run_async() and run_async_generator() dispatch to components. Done means the integration supports asynchronous pipeline execution using the new APIs, with coverage for both async entry points.