Added support for resources request in staging jobs

[andreas-kupries]

Fix #2286
Ref epinio/helm-charts#439

feat: staging job resource request support

[andreas-kupries]

Testing that the new facility operates as desired:

• Install E with --set server.stagingResourceRequests.cpu=100m --set server.stagingResourceRequests.memory=100M
• Use kubectl get pod -o yaml .... to look at the YAML of the epinio-server pod and check that the STAGING_RESOURCE_... environment variables are set to the requested values.
• Push an application and inspect the stage job pod to also have these values

Example server data

    - name: STAGING_RESOURCE_CPU
      value: 100m
    - name: STAGING_RESOURCE_MEMORY
      value: 100M

Example job data

spec:
  containers:
    - [...]
      resources:
        requests:
          cpu: 100m
          memory: 100M

[codecov]

Codecov Report

Patch coverage: 74.81% and project coverage change: +0.39 🎉

Comparison is base (fbd6f8f) 65.33% compared to head (5ffddf8) 65.72%.

❗ Current head 5ffddf8 differs from pull request most recent head 89f196b. Consider uploading reports for the commit 89f196b to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2384      +/-   ##
==========================================
+ Coverage   65.33%   65.72%   +0.39%     
==========================================
  Files         181      181              
  Lines       15412    15514     +102     
==========================================
+ Hits        10069    10197     +128     
+ Misses       4188     4168      -20     
+ Partials     1155     1149       -6     

Flag	Coverage Δ
acceptance-api	47.73% <45.18%> (+<0.01%)	⬆️
acceptance-apps	38.16% <26.66%> (-0.08%)	⬇️
acceptance-cli	62.22% <74.07%> (+0.34%)	⬆️
unittests	9.72% <11.76%> (+0.03%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
internal/api/v1/configuration/create.go	72.22% <ø> (-2.14%)	⬇️
internal/configurations/configurations.go	80.83% <0.00%> (-1.03%)	⬇️
internal/api/v1/application/deploy.go	56.66% <40.00%> (-2.27%)	⬇️
internal/api/v1/application/stage.go	75.78% <44.44%> (-0.49%)	⬇️
internal/cli/server.go	85.00% <61.53%> (-3.97%)	⬇️
internal/cli/configuration.go	66.82% <92.30%> (+8.39%)	⬆️
internal/api/v1/application/restart.go	63.26% <100.00%> (-0.74%)	⬇️
internal/api/v1/application/update.go	66.32% <100.00%> (ø)
internal/api/v1/configuration/replace.go	59.61% <100.00%> (-0.77%)	⬇️
internal/api/v1/configuration/update.go	58.00% <100.00%> (-0.83%)	⬇️
... and 6 more

... and 7 files with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[thehejik]

Works as expected.

Versions
Kubernetes Version: v1.26.5+k3s1
Epinio Server Version: v1.9.0-rc2-18-g89f196b1

Validation

• Deployed by using associated PR Expose staging job resource requests in chart to user helm-charts#439, following appended:

 --set server.stagingResourceRequests.cpu=100m --set server.stagingResourceRequests.memory=100M

• Epinio-server deployment patched to use epinio binary built from this PR, used make patch-epinio-deployment
• Checked environment of epinio-server deployment:

$ kubect get deploy  -n epinio epinio-server -o jsonpath='{..env}' | jq
...
  {
    "name": "STAGING_RESOURCE_CPU",
    "value": "100m"
  },
  {
    "name": "STAGING_RESOURCE_MEMORY",
    "value": "100M"
  }
...

• Pushed an app and checked its staging job:

$ kubectl get pod stage-workspace-sample-70bb266339423ee24e42c4426abd2f15cddrrlzs -n epinio -o yaml
...
resources:
  requests:
    cpu: 100m
    memory: 100M
...

• The deployed app itself doesn't have the resources set: resources: {}
• Input values validation check:
  • With --set server.stagingResourceRequests.cpu=100X - the helm installation fails on CrashLoopBackOff for epinio-server pod, log contains:

  ❌  bad cpu request for staging job: quantities must match the regular expression '^([+-]?[0-9.]+)([eEinumkKMGTP]*[-+]?[0-9]*)$'
  

  • With --set server.stagingResourceRequests.memory=a100hy - the helm installation fails on CrashLoopBackOff for epinio-server pod, log contains:

  ❌  bad memory request for staging job: quantities must match the regular expression '^([+-]?[0-9.]+)([eEinumkKMGTP]*[-+]?[0-9]*)$'
  

• User can set both resource types at once or just one of them

[andreas-kupries]

It should be possible to set both cpu and memory.

[thehejik]

Verified in epinio v1.9.0-rc3 according to the report above - all good