Closed sbawaska closed 5 years ago
Merging #32 into master will not change coverage. The diff coverage is
100%
.
@@ Coverage Diff @@
## master #32 +/- ##
=======================================
Coverage 84.55% 84.55%
=======================================
Files 10 10
Lines 408 408
=======================================
Hits 345 345
Misses 41 41
Partials 22 22
Impacted Files | Coverage Δ | |
---|---|---|
pkg/kab/readyness_checker.go | 89.47% <100%> (ø) |
: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 5625550...86d61df. Read the comment docs.
@glyn To me checking to see if the pod is in a desired state is exact same check kubernetes itself would perform. Since kubernetes uses exponential backoff, I used it too. Would love to hear your thoughts on why linear backoff would be better.
@glyn To me checking to see if the pod is in a desired state is exact same check kubernetes itself would perform. Since kubernetes uses exponential backoff, I used it too. Would love to hear your thoughts on why linear backoff would be better.
@sbawaska It doesn't seem unreasonable to use exponential backoff, but it might reduce responsiveness unnecessarily when it kicks in. I think it's often used to avoid making things worse when problems are caused by overloading, which might be the case here at least some of the time. Let's leave it for now and come back if responsiveness is impacted (in which case we could combine linear backoff with switching to exponential backoff eventually).
when a resource is just applied/installed, all pods may not be scheduled immediately, so be more patient.
closes #31