feat: exponential random retry

[rhajek]

Proposed Changes

Improves exponential random retry strategy in write api like in python client influxdata/influxdb-client-python#225

Delay is computed using random exponential backoff as a random value within the interval retryInterval * exponentialBase^(attempts-1) and retryInterval * exponentialBase^(attempts).

Example for retryInterval=5_000, exponentialBase=2, maxRetryDelay=125_000, total=5

Retry delays are random distributed values within the ranges of [5_000-10_000, 10_000-20_000, 20_000-40_000, 40_000-80_000, 80_000-125_000]

Checklist

• CHANGELOG.md updated
• Rebased/mergeable
• A test has been added if appropriate
• make test completes successfully
• Commit messages are in semantic format
• Sign CLA (if not already signed)

[codecov-commenter]

Codecov Report

Merging #76 (fdbb568) into master (eba35d4) will increase coverage by 0.70%.
The diff coverage is 100.00%.

@@             Coverage Diff              @@
##             master      #76      +/-   ##
============================================
+ Coverage     82.08%   82.78%   +0.70%     
- Complexity      364      375      +11     
============================================
  Files            22       23       +1     
  Lines           854      889      +35     
============================================
+ Hits            701      736      +35     
  Misses          153      153              

Impacted Files	Coverage Δ	Complexity Δ
src/InfluxDB2/DefaultApi.php	83.33% <ø> (-1.05%)	15.00 <0.00> (-2.00)
src/InfluxDB2/WriteApi.php	100.00% <100.00%> (ø)	21.00 <0.00> (-10.00)
src/InfluxDB2/WriteOptions.php	100.00% <100.00%> (ø)	1.00 <0.00> (ø)
src/InfluxDB2/WriteRetry.php	100.00% <100.00%> (ø)	17.00 <17.00> (?)

---

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 eba35d4...fdbb568. Read the comment docs.

[bednar]

Thanks for PR 👍, There is a requirement that should be satisfy before we accept the PR: