[backport v2.15.1] [Automation] - Fix imported generic cluster e2e failures

[izaac]

This is a backport of #18713 to release-2.15.

The commits applied cleanly, so the changes are the same as on master.

---

Summary

Fixes rancher/qa-tasks#2488
Fixes rancher/qa-tasks#2462
Fixes rancher/qa-tasks#2463

Occurred changes and/or fixed issues

• Fixed the whole Imported.Generic block, the three tests were failing before they reached anything they were meant to cover
• Stopped waiting on the #registration fragment, the registration tab is now selected explicitly
• Stopped gating the create test on the create form redirect, see Importing a cluster leaves you stuck on the create form with a timeout error, even though the cluster was created #18712
• Moved timeouts onto the commands that actually yield the element, several were silently inactive
• Test only changes, no product code affected

Technical notes summary

• #registration is not written by the router, it is a side effect of Tabbed picking a default tab, which happens once while no tab is active. The registration tab only exists after the registration token resolves, so if the token is slow another tab wins and the fragment never appears. A longer timeout cannot fix that ordering
• Cypress timeouts are per command, so a timeout given to rowWithName() did not carry over to column(), which fell back to the 10s default. Threaded the options through
• .should() reads its second argument as the assertion message, so timeouts passed there were ignored
• The waiting state assertion now runs before the registration manifest is applied, where no agent can have connected yet. After the apply it is a race
• The import button is now found by its own test id. It was found by position, which resolves to the create button whenever the import button is hidden
• The table of contents test now asserts against the element the dashboard actually scrolls, window.scrollY is always zero here

Testing done

The failures also needed an infrastructure fix, the imported cluster's k3s was advertising its public IP so CoreDNS could never reach the API server and the cluster never went active. That is rancher/qa-infra-automation#159.

With that in place this block goes from 3 failures to a full pass, in under a minute. Validated on this branch, the block passes 4/4 in under a minute.

Checklist

• The PR is linked to an issue and the linked issue has a Milestone, or no issue is needed
• The PR has a Milestone
• The PR template has been filled out
• The PR has been self reviewed
• The PR has a reviewer assigned
• The PR has automated tests or clear instructions for manual tests and the linked issue has appropriate QA labels, or tests are not needed
• The PR has reviewed with UX and tested in light and dark mode, or there are no UX changes
• The PR has been reviewed in terms of Accessibility
• The PR has considered, and if applicable tested with, the three Global Roles Admin, Standard User and User Base

[github-actions]

🔄 Auto-Retry Failed Run

The Tests workflow failed for the commits below and has been automatically retried.
Follow up on a retried run through its link if needed.

Commit	Workflow run	Retried at
5519ddd	31135204959	2026-08-07 01:13 UTC
8ed5d05	31135592643	2026-08-07 01:20 UTC
4d90c64	31135825961	2026-08-07 01:23 UTC