4.x: Use compressed mem tables and set a wal max entries default for quorum queues

[kjnilsson]

Compressed ETS tables may introduce a small throughput penalty (low single digit %) but can reduce peak Ra memory use by 30-50%.

Also set a default wal_max_entries value to avoid mem tables growing too large when using very small message sizes (as more than 1M tiny messages can easily fit into one WAL file).

[mkuratczyk]

Indeed, the memory savings are significant in many cases (obviosuly, mostly with long queues). The throughput difference is small and actually the branch performed a bit better in some cases - it could be caused by "test entropy" (but even then, that means the differences are small if some randomness can flip the results) but could also be actually better thanks to lower memory usage.

100B messages

In this test we hit the known issue where ra_log_segment_writer can't keep up and therefore memory usage keeps going up as new ETS tables are created before the previous are deleted.

1kb messages:
Same issue as above, but in this case it happened twice on main and once on the branch:

5kb messages:

Full results: https://grafana.lionhead.rabbitmq.com/goto/bLGbswfIR?orgId=1