Fix query for latency metrics

[Ashmita152]

Signed-off-by: Ashmita152 ashmita.bohara@logz.io

Which problem is this PR solving?

For aggregation, doing sum on latency_bucket metrics gives wrong data. We need to do sum over rate on latency_bucket

Prometheus histogram_quantile documentation also explains it.

Short description of the changes

We are replacing all sum() with sum(rate()) for latency metrics.

[codecov]

Codecov Report

Merging #3559 (314368d) into main (ab50903) will increase coverage by 0.01%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main    #3559      +/-   ##
==========================================
+ Coverage   96.48%   96.50%   +0.01%     
==========================================
  Files         263      263              
  Lines       15376    15377       +1     
==========================================
+ Hits        14836    14839       +3     
+ Misses        456      455       -1     
+ Partials       84       83       -1     

Impacted Files	Coverage Δ
plugin/metrics/prometheus/metricsstore/reader.go	100.00% <100.00%> (ø)
cmd/collector/app/server/zipkin.go	68.29% <0.00%> (-2.44%)	⬇️
cmd/query/app/static_handler.go	97.60% <0.00%> (+1.79%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update ab50903...314368d. Read the comment docs.

[albertteoh]

LGTM, thanks!