Consolidate E2E environment variables with integration varaibles

[rainest]

What this PR does / why we need it:

Source E2E test environment setup variables from the same utility package used for integration. Align E2E variables with their integration counterparts.

E2E now uses separate TEST_(KONG|CONTROLLER)_(IMAGE|TAG) variables where it previously used a single TEST_KONG(_CONTROLLER)_IMAGE_OVERRIDE variable with both the repo and tag.

CI inputs still use a single value that the CI scripts split.

Which issue this PR fixes:

Fix #4593

Special notes for your reviewer:

GKE-related variables are defined by the related KTF module, and have not been changed.

License data was also using a KTF-defined variable from the Kong addon, although it isn't actually using the Kong addon. Integration does use the same var AFAIK, however, since it does use the Kong addon. I left this unchanged.

https://github.com/rainest/kubernetes-ingress-controller/actions/runs/6305128682/job/17117913357 and https://github.com/rainest/kubernetes-ingress-controller/actions/runs/6305130237/job/17119050768 have debug-enabled CI runs on a fork. GKE and Enterprise license failures are expected for the fork.

This is mostly to confirm the split script is populating the expected variables, since those variable names actually changed in E2E:

2023-09-25T22:23:58.5265037Z   TEST_CONTROLLER_IMAGE: kong/kubernetes-ingress-controller
2023-09-25T22:23:58.5265441Z   TEST_CONTROLLER_TAG: sha-d140eb7
2023-09-25T22:23:58.5265766Z   TEST_KONG_LOAD_IMAGES: true
2023-09-25T22:23:58.5266116Z   TEST_KONG_IMAGE: kong/kong-gateway-dev
2023-09-25T22:23:58.5266440Z   TEST_KONG_TAG: nightly

Other variables just got redefined at a new location.

PR Readiness Checklist:

Complete these before marking the PR as ready to review:

• the CHANGELOG.md release notes have been updated to reflect any significant (and particularly user-facing) changes introduced by this PR

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (93c1a9c) 77.6% compared to head (dca814a) 77.8%.
Report is 23 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##            main   #4727     +/-   ##
=======================================
+ Coverage   77.6%   77.8%   +0.1%     
=======================================
  Files        164     165      +1     
  Lines      18453   18477     +24     
=======================================
+ Hits       14331   14386     +55     
+ Misses      3299    3273     -26     
+ Partials     823     818      -5     

see 24 files with indirect coverage changes

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[pmalek]

One nit and conflicts to resolve.

[team-k8s-bot]

E2E (targeted) tests with KIND-based clusters were started at https://github.com/Kong/kubernetes-ingress-controller/actions/runs/6566314300

[rainest]

More conflicts. #4766 removed some generic E2E utilities that weren't specific to <3.4 gates. While we chose a required version newer than any of the extant gates now, we probably will have similar gates in the future, so keeping the generic functions around with nolint:unused.

[randmonkey]

The nolint:unused comment did not take its effect. Is there any difference between the local make lint and lint job on the CI ?

[pmalek]

In order to prevent the linter fail you'd need to add

, "-U1000" to

kubernetes-ingress-controller/staticcheck.conf

Line 2 in 1fa567f

checks = ["all", "-ST1000", "-ST1005"]

We run both staticcheck and golangci-lint as part of lint and they don't take into account the same code comment markers.

[pmalek]

istio tests failed in e2e suite: https://github.com/Kong/kubernetes-ingress-controller/actions/runs/6566314300/job/17836886646

[team-k8s-bot]

E2E (targeted) tests with KIND-based clusters were started at https://github.com/Kong/kubernetes-ingress-controller/actions/runs/6592767672

[rainest]

https://github.com/Kong/kubernetes-ingress-controller/actions/runs/6592767672/job/17914168070 Istio run now happy. Do we have a way to run GKE tests off a branch? It hopefully just needs the same thing with a static controller image (I'd neglected to add the new input splitter to both it and Istio), but https://github.com/Kong/kubernetes-ingress-controller/actions/runs/6592767672 just skips it.

It shouldn't have additional complications given that it's just the main E2E job with GKE flipped on, but if we can run it, may as well.

Edit: https://github.com/Kong/kubernetes-ingress-controller/actions/runs/6592883775 the nightly job has a branch dispatch option, so that should get it and anything else that differs from the label variant.