Refactor kafka_int_tests to use connection creator helper

[rukai]

The driving factor for this was the ability to abstract over connections created with TLS, TLS+SASL,plaintext+SASL or just plaintext.

In the future I expect this to also be useful for abstracting over different kafka drivers.

The consumer and producer APIs are completely abstracted and the tests no longer directly interact with rdkafka.
However the rdkafka admin API is still directly exposed since that has a lot more surface area.
We can wrap that up if and when we need to abstract over multiple kafka drivers.

[github-actions]

2 benchmark regressed. 0 benchmark improved. Please check the benchmark workflow logs for full details: https://github.com/shotover/shotover-proxy/actions/runs/8000182500

                        change: [+5.8377% +22.807% +51.668%] (p = 0.04 < 0.05)
                        Change within noise threshold.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high severe
kafka_codec/encode_all  time:   [2.5030 µs 2.5451 µs 2.5862 µs]
                        change: [+29.058% +35.012% +40.893%] (p = 0.00 < 0.05)
                        Performance has regressed.
--
  3 (3.00%) low mild
  6 (6.00%) high mild
  5 (5.00%) high severe
cassandra_codec/encode_system.local_result_v4_no_compression
                        time:   [32.973 µs 34.034 µs 34.994 µs]
                        change: [+26.121% +31.687% +37.149%] (p = 0.00 < 0.05)
                        Performance has regressed.

[github-actions]

1 benchmark regressed. 0 benchmark improved. Please check the benchmark workflow logs for full details: https://github.com/shotover/shotover-proxy/actions/runs/8010223586

  2 (2.00%) low mild
  1 (1.00%) high mild
  1 (1.00%) high severe
kafka_codec/decode_request_produce_drop
                        time:   [491.73 ns 523.86 ns 559.20 ns]
                        change: [+27.341% +33.654% +40.816%] (p = 0.00 < 0.05)
                        Performance has regressed.

[github-actions]

1 benchmark regressed. 0 benchmark improved. Please check the benchmark workflow logs for full details: https://github.com/shotover/shotover-proxy/actions/runs/8011642761

  1 (1.00%) low mild
  1 (1.00%) high mild
  3 (3.00%) high severe
kafka_codec/encode_request_list_offsets
                        time:   [762.93 ns 770.68 ns 777.65 ns]
                        change: [+21.384% +53.449% +110.44%] (p = 0.02 < 0.05)
                        Performance has regressed.