2
62%
master: %

Ran 08 Jan 2019 09:55PM UTC

Files 100

Run time 22s

Badge

Embed ▾

Committed 08 Jan 2019 09:35PM UTC coverage: 61.049% (+0.5%) from 60.542%

Job # LATEST_GO=true

Build Type

push

travis-ci

Committed by

buger

Commit Message

Fix host checker (#2049)

Current implementation has multiple flaws. 
At the moment first HostDown event happens after a number of specified tries, and if the host is still down, the next HostDown event will happen after the same number of tries. So if `time_wait` is 10s, and `failure_trigger_sample_size` 2, it means 20 seconds between events. At the same time we set redis expiration key for "host down" key, to `time_wait` + 1, which means key will be expired 9 seconds before second HostDown event. This PR fix it by setting expiration time to `time_wait` * `failure_trigger_sample_size`

Another issue is how host considered to be up. At the moment on first succesful attempt we just enabled the host, however if upstream is unstable, having single healthy attempt does not mean that it is recovered. This change makes Host UP logic work exactly like Host Down and now consider number of tries. 

Fix https://github.com/TykTechnologies/tyk/issues/2036

Coverage Stats

9700 of 15889 relevant lines covered (61.05%)

0.68 hits per line

TykTechnologies / tyk / 5947 / 2
62%
master: %

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job 5947.2 (LATEST_GO=true)

TykTechnologies / tyk / 5947 / 2 62% master: %

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job 5947.2 (LATEST_GO=true)

TykTechnologies / tyk / 5947 / 2
62%
master: %

README BADGES
x