TBS: Optimize for ReadTraceEvents miss

[carsonip]

Motivation/summary

When a sampling decision is received from a remote apm-server, if the trace is not stored locally, prefetching values in the iterator is causing mmap vlog to be read for no value, resulting in high memory usage and disk read IO.

Add an initial pass with PrefetchValues=false to optimize for a miss in ReadTraceEvents.

e.g. if there are 10 apm-servers all running TBS and synced via ES, and all transactions/spans from a trace are always sent to in 1 apm-server (i.e., a single trace never stored across multiple apm-servers), this PR should cut disk read IO related to sampling decision handling to 1/10 of before.

Benchmark results

Since the benchmark runs badger DB with in-memory mode, I expect in reality, hit=false after this fix will stay the same, while all the other numbers to be 1 order of magnitude slower. i.e. the performance gain from this PR will be even greater.

goos: linux
goarch: amd64
pkg: github.com/elastic/apm-server/x-pack/apm-server/sampling/eventstorage
cpu: 11th Gen Intel(R) Core(TM) i7-11800H @ 2.30GHz
                                       │ bench-read-events-hit.before │     bench-read-events-hit.after     │
                                       │            sec/op            │    sec/op     vs base               │
ReadEventsHit/bigTX=true/hit=false-16                   167.06µ ± 14%   14.31µ ± 32%  -91.43% (p=0.002 n=6)
ReadEventsHit/bigTX=true/hit=true-16                     174.4µ ±  4%   186.3µ ±  4%   +6.81% (p=0.004 n=6)
ReadEventsHit/bigTX=false/hit=false-16                  129.49µ ±  3%   16.93µ ± 21%  -86.92% (p=0.002 n=6)
ReadEventsHit/bigTX=false/hit=true-16                    136.5µ ±  3%   139.8µ ±  3%   +2.40% (p=0.026 n=6)
geomean                                                  150.7µ         50.13µ        -66.73%

Checklist

- [ ] Update CHANGELOG.asciidoc Require backport release changelog
- [ ] Documentation has been updated

For functional changes, consider:

• Is it observable through the addition of either logging or metrics?
• Is its use being published in telemetry to enable product improvement?
• Have system tests been added to avoid regression?

How to test these changes

See benchmark

Related issues

Closes #13463