Fix prometheus histogram metric tracker for multiple pools

[vaIgarashi]

Currently it's not possible to use PrometheusMetricsTrackerFactory if you have multiple pools. Even if you create one instance of factory it's not thread-safe and collector created multiple times.

Related to #1331
Fixes #1465

[codecov]

Codecov Report

Merging #1692 (2220196) into dev (8217f20) will not change coverage.
The diff coverage is n/a.

@@            Coverage Diff            @@
##                dev    #1692   +/-   ##
=========================================
  Coverage     70.11%   70.11%           
  Complexity      561      561           
=========================================
  Files            26       26           
  Lines          2118     2118           
  Branches        296      296           
=========================================
  Hits           1485     1485           
  Misses          487      487           
  Partials        146      146           

---

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 8217f20...2220196. Read the comment docs.

[rabindrarakshit]

@brettwooldridge Could you please check this. It will be good to have this in. We want to use the Histogram approach, the normal summary approach blocks a huge number of threads in prometheus code blocks for a new pod scaled during peak hours.

Also this looks duplicate of these: #1465, #1467
May be we can close one.

[brettwooldridge]

@rabindrarakshit Ok, your chance to play God. Which of the two pull requests is better? Better being: concise and correct with good test coverage.

[rabindrarakshit]

Thanks for the honour @brettwooldridge :) I would like to go with this : #1692 It looks good to the point (also the ability to remove metrics when connection pool is shutting down) except one minor remark about the .* imports. If you are fine with that then we can proceed with this.

Both are somewhat copy paste of this : #1331

[rabindrarakshit]

@brettwooldridge could you please approve this pr so that we can get it merged and possibly a release would be very good. We want to test the histogram and see if it brings any improvement over the classical summary approach!

[brettwooldridge]

@rabindrarakshit I intend to release within the next week. Trying to get a few more pull requests integrated.

[rabindrarakshit]

ok that would be great. Thanks @brettwooldridge