feat: Implement ReadN

[raulb]

Description

Since the introduction of end-to-end batching on the Connector SDK, (part of the Connector SDK 0.12.0), source connectors can add support to batching by implementing the ReadN method. This wasn't a breaking change since the SDK will fallback to the traditional Read method as it can be spotted in the logs when starting Conduit.

Example:

2025-04-24T19:21:21+00:00 INF source does not support batch reads, falling back to single reads component=plugin connector_id=postgres-to-kafka:postgres-source plugin_name=builtin:postgres plugin_type=source

This pull-request implements ReadN while still maintaining support to Read method for those who wish to use this connector in an older version of Conduit. This is probably not necessary considering Postgres is a builtin connector.

To really utilize the advantage of this, it is recommended to run Conduit with the flag that enables a new upcoming architecture via --preview.pipeline-arch-v2. More information on this blog post.

Note

Still need to:

• Remove Read and Next leaving only ReadN and NextN

Running benchmarks

Inserting 20M records using Benchi using CDC:

• WIth Read: 142824.47 msg/s
• With ReadN: 153107.18 msg/s (previously )

This is a 7,2% better on CDC.

Quick checks:

• I have followed the Code Guidelines.
• There is no other pull request for the same update/change.
• I have written unit tests.
• I have made sure that the PR is of reasonable size and can be easily reviewed.