optimizely / go-sdk / 2057 / 4
92%
master: 92%

DEFAULT BRANCH: master
Ran
Files 49
Run time 4s
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 91.276% (+0.03%) from 91.244%
GIMME_GO_VERSION=1.12.x

push

travis-ci-com

web-flow 
(chore): benchmark different event processor configurations (#187)

* bump the queue size to 2k

* benchmark with 0 events thrown out and a throughput of 1000 calls at 1.3 ms per call with a 1 ms sleep.  looks like about .3 ms per call

* add some benchmarking

* test benchmarking for fails or no fails

* make the customer logger buffer thread safe via a queue

* 200 ms sleep between calls to allow go routines to run.

* take out sleep for processing

* resolve race condition

* terminate processor

* refactor tests

* don't print out benchmark logs

* default failing

* print the count of lost events

* tweaking to get tests to pass

* check 100, 2000, and large 4000 queue

* try testing different queue sizes

* add locking around global default logging consumer

* how big does the queue need to be?

* try 100,1000,10000 as queue sizes

* get processor batch tests working with benchmarks

* tweaking the benchmarking

* tweaking sleep interval to 1 nanosecond

* moving sleep

* moving that 1 nanosecond sleep around to see if it makes a difference.  I assume once the queue gets a hold, it is drained.

* one more refactor

* refactor and show that percent loss can be brought down with larger queue

* add a bunch of benchmarking

* cleanup benchmarking

* fix spacing

* refactor tests to continue to call processEvent until the event is processed

* fix processor to pass back false on fail

* fix client test

* take out unused fun

* add benchmark tests to CI and makefile

* take out datadog and yaml mod and sum updates

2260 of 2476 relevant lines covered (91.28%)

8342.44 hits per line

Source Files on job 2057.4 (GIMME_GO_VERSION=1.12.x)