[aggregator] Sort heap in one go, instead of iterating one-by-one

[vdarulis]

What this PR does / why we need it:

A few more small timer aggregation tweaks.
Total throughput with 12 concurrent streams:

name                                                              old speed      new speed      delta
TimerAddBatch/100k_samples_1000_batch_size-12                      604MB/s ± 3%   633MB/s ± 0%   +4.84%  (p=0.008 n=5+5)
TimerAddBatch/100k_samples_1000_batch_size_0..1,000,000_range-12   555MB/s ± 5%   635MB/s ± 0%  +14.50%  (p=0.008 n=5+5)
TimerAddBatch/100k_samples_1000_batch_size_0..100,000_range-12     567MB/s ± 1%   584MB/s ± 5%     ~     (p=0.151 n=5+5)
TimerAddBatch/10k_samples_1000_batch_size-12                       543MB/s ± 2%   590MB/s ± 2%   +8.54%  (p=0.008 n=5+5)
TimerAddBatch/10k_samples_100_batch_size-12                        536MB/s ± 1%   583MB/s ± 2%   +8.83%  (p=0.008 n=5+5)
TimerAddBatch/10k_samples_100_batch_size_with_negative_values-12   537MB/s ± 1%   575MB/s ± 2%   +7.09%  (p=0.008 n=5+5)
TimerAddBatch/1k_samples_100_batch_size-12                         621MB/s ± 1%   652MB/s ± 2%   +4.92%  (p=0.008 n=5+5)

Single CPU only:

name                                                           old speed      new speed      delta
TimerAddBatch/100k_samples_1000_batch_size                     74.2MB/s ± 0%  80.0MB/s ± 0%  +7.74%  (p=0.008 n=5+5)
TimerAddBatch/100k_samples_1000_batch_size_0..1,000,000_range  74.3MB/s ± 0%  79.6MB/s ± 1%  +7.17%  (p=0.008 n=5+5)
TimerAddBatch/100k_samples_1000_batch_size_0..100,000_range    74.1MB/s ± 0%  79.9MB/s ± 1%  +7.87%  (p=0.008 n=5+5)
TimerAddBatch/10k_samples_1000_batch_size                      75.8MB/s ± 0%  80.9MB/s ± 1%  +6.78%  (p=0.008 n=5+5)
TimerAddBatch/10k_samples_100_batch_size                       76.0MB/s ± 1%  81.5MB/s ± 0%  +7.22%  (p=0.008 n=5+5)
TimerAddBatch/10k_samples_100_batch_size_with_negative_values  75.6MB/s ± 0%  81.2MB/s ± 1%  +7.42%  (p=0.008 n=5+5)
TimerAddBatch/1k_samples_100_batch_size                        91.7MB/s ± 0%  97.1MB/s ± 1%  +5.80%  (p=0.008 n=5+5)

Special notes for your reviewer:

Does this PR introduce a user-facing and/or backwards incompatible change?:

Does this PR require updating code package or user-facing documentation?:

[codecov]

Codecov Report

Merging #3331 (e995ae4) into master (a03e55f) will increase coverage by 26.8%.
The diff coverage is 98.7%.

@@             Coverage Diff             @@
##           master    #3331       +/-   ##
===========================================
+ Coverage    45.6%    72.5%    +26.8%     
===========================================
  Files           4     1099     +1095     
  Lines         274   101668   +101394     
===========================================
+ Hits          125    73727    +73602     
- Misses        139    22864    +22725     
- Partials       10     5077     +5067     

Flag	Coverage Δ
aggregator	76.6% <98.7%> (?)
cluster	84.9% <ø> (?)
collector	84.3% <ø> (+38.6%)	⬆️
dbnode	78.9% <ø> (?)
m3em	74.4% <ø> (?)
m3ninx	73.5% <ø> (?)
metrics	20.0% <ø> (?)
msg	74.2% <ø> (?)
query	67.4% <ø> (?)
x	80.4% <ø> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

---

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 a03e55f...caf401f. Read the comment docs.