Validate: Remove Long Running `test-on-omnibus` Pipeline Setup for Backport MRs

Problem

Validate if e2e:test-on-omnibus-ee is executed on backport MRs, and transition them if so.

End-to-end suites

e2e:test-on-omnibus-ee full suite in tier 3

qa:test-on-gdk gdk-instance suite in tier 3

Omnibus and CNG pipelines have significantly different runtimes

⏲️ ~~~95 minutes saved per MR pipeline on its critical path (~140 mins omnibus -> ~45 mins cng Dashboard)~~
💵 ~~~10'000USD/month saved (calculation)~~
🧱 Less pipeline false failures due to ability to block MRs based on test failures. Previously this gate was loosely added.

To save time, money, and false failures, transition any e2e:test-on-omnibus type pipelines to e2e:test-on-cng if possible.

Baseline e2e:test-on-omnibus-ee as of 2024-05-02:

Tier 1 (ref): ~95m Tier 3 (ref): ~95m

Remove e2e:test-on-omnibus-ee from Tier 1 pipelines. (and backport to past 2 releases)
Change e2e:test-on-omnibus-ee Tier 3 to run smoke-only and orchestrated scenarios. (and backport to past 2 releases) - Change `e2e:test-on-omnibus-ee` Tier-3 to run s... (#21285 - closed)
Replace e2e:test-on-gdk to use e2e:test-on-cng on Tier 3 pipelines to run instance scenarios.(and backport to past 2 releases) - Replace e2e:test-on-gdk to use e2e:test-on-cng ... (#21286 - closed)

Task	Result
Removed `e2e:test-on-omnibus-ee` from tier-1 backport MR pipelines	Reduces tier 1 pipeline runtime ~40% without compromising quality.
Removed `instance` test suite (main e2e test suite) from `test-on-omnibus`	Reduces costs by not running same test suite against several deployment types, reduces flaky failure probability
Removed `test-on-gdk` and replaced it with `test-on-cng`	Higher confidence in patch release as it now uses actual customer like deployment instead of dev deployment, P80 feedback time improvement by 20 minutes which is the difference between `test-on-cng` and `test-on-gdk` pipeline runtimes

Edited Jul 24, 2025 by andrey