[Minor] Backport changes to metadata benchmark

[etseidl]

Which issue does this PR close?

• Part of Use custom thrift decoder to improve speed of parsing parquet metadata #5854.

Rationale for this change

Backport changes to allow apples-to-apples comparison of thrift decoding

What changes are included in this PR?

Adds a page header benchmark and updates bench names to match those in feature branch.

Are these changes tested?

No tests needed...only changes to benchmark

Are there any user-facing changes?

No

[etseidl]

With these changes, I can again use critcmp

% critcmp main read_page_header
group                                 main                                   read_page_header
-----                                 ----                                   ----------------
decode parquet metadata               1.87     35.8±0.98µs        ? ?/sec    1.00     19.1±0.33µs        ? ?/sec
decode parquet metadata (wide)        2.73    221.4±2.90ms        ? ?/sec    1.00     81.0±2.87ms        ? ?/sec
decode thrift file metadata           2.22     21.8±0.31µs        ? ?/sec    1.00      9.8±0.20µs        ? ?/sec
decode thrift file metadata (wide)    1.80    112.7±2.39ms        ? ?/sec    1.00     62.8±1.27ms        ? ?/sec
open(default)                         1.79     36.2±1.00µs        ? ?/sec    1.00     20.2±0.49µs        ? ?/sec
open(page index)                      6.80  1773.7±23.97µs        ? ?/sec    1.00    260.7±4.39µs        ? ?/sec
page headers                          1.67     12.1±0.27µs        ? ?/sec    1.00      7.3±0.17µs        ? ?/sec
page headers (no stats)                                                      1.00      4.2±0.06µs        ? ?/sec

[etseidl]

Adding some encoding stats because profiling showed around 25% of the time to read the file metadata struct is spent reading this vector 😮

[etseidl]

Thanks @mbrobbel!