Profile a query using Microlog

### Feature Request / Improvement

At PyData Amsterdam 2025 we had an awesome talk on Microlog:

I'm more execited about this than I should, and curious to learn what it thinks of the time spend on a typical query.

It would be good to run this on the benchmark: https://github.com/apache/iceberg-python/blob/main/tests/benchmark/test_benchmark.py

Thanks @laffra 🙌

Start with tests/benchmark/test_benchmark.py and review how the existing benchmark runs a typical query. Investigate how Microlog can profile that benchmark without changing what it measures. Done means the benchmark can be run with Microlog and produces useful timing information for the query.

Written by the indexing model from the issue text.