Closed omus closed 3 years ago
Merging #34 (8b49288) into main (a605bbe) will increase coverage by
16.79%
. The diff coverage is72.00%
.
@@ Coverage Diff @@
## main #34 +/- ##
===========================================
+ Coverage 34.67% 51.47% +16.79%
===========================================
Files 4 4
Lines 124 136 +12
===========================================
+ Hits 43 70 +27
+ Misses 81 66 -15
Impacted Files | Coverage Δ | |
---|---|---|
src/K8sClusterManagers.jl | 100.00% <ø> (+18.18%) |
:arrow_up: |
src/native_driver.jl | 4.41% <7.14%> (-0.07%) |
:arrow_down: |
src/pod.jl | 97.91% <97.22%> (+47.91%) |
: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 a605bbe...8b49288. Read the comment docs.
Marking as breaking as this requires users to update their configure
functions
Using Kuber.jl seems like more trouble than it's worth at the moment.
getfield
andsetfield
access requiring a work around when non-valid field strings are required. e.g. the updated GPU configuration function mentioned in the README would need to be written as:kube_obj
requires a JSON string orDict
. It cannot acceptAbstractDict
.kubectl proxy
commandKuberContext
has been problematic in the CI environment (https://github.com/beacon-biosignals/K8sClusterManagers.jl/pull/31#issuecomment-824087436)Update: Additional reasons to get rid of Kuber.jl:
kubectl proxy
can cause exceptions to be raised which require requests to be retried. Such failures can also cause thekubectl proxy
to crash which results in the user needing to know how to restart the proxy command