perf(search): replace map with slice to improve histogram performance

[eguguchkin]

replace map with slice to improve performance for billion-access queries

Fixes #170

---

• I have read and followed all requirements in CONTRIBUTING.md;
• I used LLM/AI assistance to make this pull request;

[codecov-commenter]

Codecov Report

❌ Patch coverage is 81.81818% with 6 lines in your changes missing coverage. Please review.
✅ Project coverage is 71.58%. Comparing base (abd6bfb) to head (5251276).
⚠️ Report is 4 commits behind head on main.

Files with missing lines	Patch %	Lines
frac/processor/search.go	79.31%	5 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #171      +/-   ##
==========================================
+ Coverage   71.37%   71.58%   +0.20%     
==========================================
  Files         200      200              
  Lines       18191    18216      +25     
==========================================
+ Hits        12984    13040      +56     
+ Misses       4492     4458      -34     
- Partials      715      718       +3     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[cheb0]

Looks good. However, it seems performance is not for free, because this hist can be quite big if user sets a low hist interval like 1s (or even lower). In that case, hists will take almost 700 kbytes of memory for 24h fraction. If number of hit buckets is low, map will probably win, because updating small memory set (which fit to L1 cache) is more efficient even despite of map.put call overhead.

However, with map it might be even worse if hist interval is low and every bucket has hits. In such scenarios, we run build histograms, but they can not be sent by GRPC (size limit).