refactor: Refactor snowflake to use spmc abstractions

[tomasfarias]

Problem

Refactor Snowflake batch export to use fully-async SPMC abstractions. Big part of the work was done in previous PRs, so there isn't much to do here.

Changes

1. Implement new SnowflakeConsumer class
2. Move flush function to flush method of SnowflakeConsumer.
3. Call SnowflakeConsumer with run_consumer_loop in S3 batch export.
4. Remove file_no from SnowflakeHeartbeatDetails (and associated tests). Since every consumer writes one file, we don't have a need to add a file_no to each one. Every file name is unique as NamedTemporaryFile generates a new name for each one.

TODO:

• Wondering if we could use parquet here too?

👉 Stay up-to-date with PostHog coding conventions for a smoother review.

Does this work well for both Cloud and self-hosted?

How did you test this code?

Manually ran Snowflake unit tests, and everything passed (with the exception of the two tests that had to be fixed.