Conversation
Using GenericEvents and a channel produces inconsistent results. Most OSS operators don't implement this functionality. Switching back to creating the gitopsconfig job directly. (Which is the same the rconciler would do, except this works consistently)
Codecov Report
@@ Coverage Diff @@
## master #354 +/- ##
==========================================
+ Coverage 24.81% 25.00% +0.18%
==========================================
Files 6 6
Lines 540 536 -4
==========================================
Hits 134 134
+ Misses 377 373 -4
Partials 29 29
Continue to review full report at Codecov.
|
| echo "Error: Argument for $1 is missing" >&2 | ||
| exit 1 | ||
| fi | ||
| ;; |
There was a problem hiding this comment.
Curious as to why this is done via environment variables vs managing the settings in the code?
There was a problem hiding this comment.
It's about the scope of the variables. Using export allows us to be more flexible within the shell script. The env variables never make it to the parent shell, so there's no concern about it changing anything important.
We also need to use environment variables so that we can control how make behaves when called from .travis.yml
3 participants
Description
This pull request fixes the issue with unreliable webhook triggers.
Using GenericEvents and a channel produces inconsistent results. Most OSS operators don't implement this functionality. Switching back to creating the gitopsconfig job directly (Which is the same the reconciler would do, except this works consistently).
Fixes #336
This PR also adds 2 new parameters to
scripts/e2e-test.shin order to speed up local testing cycles:Type of change
Checklist