server/*: add trend to check etcd healthy

[HuSharp]

What problem does this PR solve?

Issue Number: Ref #7251

What is changed and how does it work?

• introduce the trend struct which has 2 windows, related to slow node detection
  • small windows duration is 20s, big windows duration is 60s
  • rate = (small_windows_avg - big_windows_avg - unSenseVal)/small_windows_avg
  • rate indicates the input's changing rate
  • The average value is from fsync which is recorded in etcd's server side
  • unSenseVal related to the environment, which is calculated by all average fsync val
• event been triggered by 2 aspects: etcd leader changed & rate less than point
  • put the check rate at the campaign leader
  • record rate every 1 seconds
• We have 4 formulas:
  • (A-B)/B >= m, (rate)
  • Sa = Ax, Sb = nBx, (n is a multiple of window big for window small, x is small window duration)
  • Sb = δ + Sa
  • δ = (big_dur-small_dur)*base (base determined by environment)
    We can get -> δ/Sb = δ/(δ+Sa) <= (m+1-n) => (big_dur-small_dur) * base / (base * bid_dur +spike * t) <= (m+1-n)
• for now, we want to monitor spike/base = 50, which means needs to grow 100 times what before
  • we can get: (big_dur-small_dur)b / (small_durb + 100 * t * b) <= (m+1-n)/n
  • for now: big_dur=60, small_dur=20, n=3, s=50, and spike times be 3: 40b/(60b+50 * 3 * b) <= (m-2) -> m=2.19
  • I set n = 3. and when n is fixed increasing m means increasing the difficulty of triggering

rate graph

fsync graph

Check List

Tests

• Unit test
• Integration test
• Manual test (add detailed scripts or steps below)

Release note

None.

[ti-chi-bot]

[REVIEW NOTIFICATION]

This pull request has not been approved.

To complete the pull request process, please ask the reviewers in the list to review by filling /cc @reviewer in the comment.
After your PR has acquired the required number of LGTMs, you can assign this pull request to the committer in the list by filling /assign @committer in the comment to help you merge this pull request.

The full list of commands accepted by this bot can be found here.

Reviewer can indicate their review by submitting an approval review.
Reviewer can cancel approval by submitting a request changes review.

[ti-chi-bot]

Skipping CI for Draft Pull Request.
If you want CI signal for your change, please convert it to an actual PR.
You can still manually trigger a test run with /test all

[codecov]

Codecov Report

Merging #7682 (65f5f51) into master (a90e13e) will decrease coverage by 0.06%.
Report is 2 commits behind head on master.
The diff coverage is 82.95%.

❗ Current head 65f5f51 differs from pull request most recent head de02922. Consider uploading reports for the commit de02922 to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #7682      +/-   ##
==========================================
- Coverage   73.82%   73.77%   -0.06%     
==========================================
  Files         429      432       +3     
  Lines       47542    47668     +126     
==========================================
+ Hits        35100    35165      +65     
- Misses       9458     9492      +34     
- Partials     2984     3011      +27     

Flag	Coverage Δ
unittests	73.77% <82.95%> (-0.06%)	⬇️

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

[HuSharp]

Fixed by #7737.