Benchmark MaxScale 2.3 using the same setup as in
MXS-1597, so that the results are comparable.
Use the tag `maxscale-2.3.0`.
Please run the benchmarks both with query_classifier_cache_size=0 defined in the [MaxScale] section, and with that configuration setting not defined at all.
In the former case, the query classification will be performed exactly like in 2.2, so the benchmark results will thus show how 2.3 compares with 2.2, with the general mechanisms being roughly the same. In the latter case, the query classification results will be cached and reused, which should lead to better performance.