Create matching benchmarks for the arrow datastore

[jleibs]

If these look terrifying, don't panic. Improvements are coming in a followup PR:

obj_mono_points/insert  time:   [107.87 ms 108.56 ms 109.41 ms]
                        thrpt:  [91.400 Kelem/s 92.118 Kelem/s 92.700 Kelem/s]
                 change:
                        time:   [+5529.6% +5637.2% +5741.8%] (p = 0.00 < 0.05)
                        thrpt:  [-98.288% -98.257% -98.224%]
                        Performance has regressed.

obj_mono_points/query   time:   [21.120 ms 21.316 ms 21.533 ms]
                        thrpt:  [4.6440 Kelem/s 4.6914 Kelem/s 4.7348 Kelem/s]
                 change:
                        time:   [+229611% +232068% +234698%] (p = 0.00 < 0.05)
                        thrpt:  [-99.957% -99.957% -99.956%]
                        Performance has regressed.

obj_batch_points/insert time:   [492.58 µs 493.57 µs 494.80 µs]
                        thrpt:  [20.210 Melem/s 20.261 Melem/s 20.301 Melem/s]
                 change:
                        time:   [+31.371% +31.914% +32.426%] (p = 0.00 < 0.05)
                        thrpt:  [-24.486% -24.193% -23.880%]
                        Performance has regressed.

obj_batch_points/query  time:   [285.54 µs 290.11 µs 295.94 µs]
                        thrpt:  [337.91 Kelem/s 344.70 Kelem/s 350.22 Kelem/s]
                 change:
                        time:   [+43136% +44278% +45463%] (p = 0.00 < 0.05)
                        thrpt:  [-99.781% -99.775% -99.769%]
                        Performance has regressed.

Resolves: #575

Checklist

• I have read and agree to Contributor Guide and the Code of Conduct
• I've included a screenshot or gif (if applicable)
• I've added a line to CHANGELOG.md (if this is a big enough change to warrant it)

[emilk]

Awesome to finally get these in!

[emilk]

in crates/re_data_store/benches/obj_query_benchmark.rs we use 1000 for both these. I guess arrow is still too slow for that?

It would be nice to have matching numbers so we can compare apples to apples

[jleibs]

Yup -- this is coming as soon as we land #582

[emilk]

unrelated, but message encoding/decoding is another thing we would like to benchmark!