feat(replays): Add buffered consumer implementation

[cmanallen]

TODO description...

Partially addresses the DACI: https://www.notion.so/sentry/DACI-Session-Replay-Recording-Consumer-Stability-and-Performance-Improvements-19e8b10e4b5d80a192a1ecd46f13eebb

How it works:

• As messages come in they are processed and their processed results are stored on a queue.
• When the queue fills up the processed messages are flushed.
  • Flushing involves committing data to GCS, ClickHouse, BigQuery, DataDog.
  • Anything I/O related.
• Flushing happens in a thread-pool.

This closely mirrors current production behavior except processing is no longer done in the thread-pool.

This PR also introduces a new RunTime abstraction for managing state changes in the consumer. Which I will document in a Notion doc.

[codecov]

Codecov Report

Attention: Patch coverage is 98.68132% with 6 lines in your changes missing coverage. Please review.

✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
src/sentry/replays/consumers/buffered/platform.py	96.36%	4 Missing ⚠️
src/sentry/replays/consumers/buffered/consumer.py	98.03%	1 Missing ⚠️
...ests/sentry/replays/unit/consumers/test_runtime.py	98.00%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff            @@
##           master   #85356    +/-   ##
========================================
  Coverage   87.89%   87.89%            
========================================
  Files        9701     9711    +10     
  Lines      550004   550638   +634     
  Branches    21376    21376            
========================================
+ Hits       483412   484007   +595     
- Misses      66283    66322    +39     
  Partials      309      309