scherniavsky commented 5 years ago

This PR does 2 things:

addresses https://github.com/zalando-incubator/docker-locust/issues/108 (fail on wrong # of slaves)
minimizes this issue when master / slave startup order can't be controlled

It was observed that if slaves start shortly before a master, it takes them much time to reconnect, which:

sometimes exceeds slaves waiting timeout
may lead to higher load if slaves connect to the master during a running test

The PR adds master readiness probe on slaves, so they start after a master and don't need to reconnect.

codecov-io commented 5 years ago

Codecov Report

Merging #109 into master will decrease coverage by 0.12%. The diff coverage is 91.11%.

@@            Coverage Diff             @@
##           master     #109      +/-   ##
==========================================
- Coverage    91.8%   91.68%   -0.13%     
==========================================
  Files           7        7              
  Lines         525      553      +28     
==========================================
+ Hits          482      507      +25     
- Misses         43       46       +3

Impacted Files	Coverage Δ
src/tests/test_bootstrap.py	`96.06% <81.81%> (-3.07%)`	:arrow_down:
src/app.py	`88.65% <94.11%> (+1.26%)`	:arrow_up:

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 1f19a41...12aa623. Read the comment docs.

scherniavsky commented 5 years ago

:+1:

thilp commented 5 years ago

:+1:

scherniavsky commented 5 years ago

:+1:

scherniavsky commented 5 years ago

@tortila after we agreed not to implement the additional changes I've tried to resolve your conversations, but they still block the PR from being merged. Could you fix them please?

zalando-incubator / docker-locust

Fail on wrong # of slaves, probe master readiness #109

Codecov Report