refactor(e2e): support stopping on failure

[diafour]

Description

• STOP_ON_FAILURE=yes env to stop test on first failure
• Taskfile changes:
  • Rename run to run:ci, run_local to run.
  • Remove dedicated task run_one.
  • Get Ginkgo version from go.mod, remove hardcoded envs from workflow and Taskfile.
  • Change "ginkgo" task from install to check behaviour: warn about ginkgo absence and version mismatch.
• Restructure e2e documentation:
  • Restructure.
  • Add documentation for FOCUS and STOP_ON_FAILURE env.
  • Add cluster connection settings priority.

Why do we need it, and what problem does it solve?

• Run tests without cleanup for debugging purposes.
• More consistent configuration.

What is the expected result?

• I can see resources in cluster in erroneous state if error occurs in test.

FOCUS env effect:

$ task run
ginkgo -v --skip-file vm_test.go --skip-file vm_label_annotation_test.go --skip-file ipam_test.go --skip-file disks_test.go

$ FOCUS="Complex test" task run
ginkgo -v --focus Complex test

$ task run:ci                  
ginkgo --skip-file vm_test.go --skip-file vm_label_annotation_test.go --skip-file ipam_test.go --skip-file disks_test.go -v

Checklist

• The code is covered by unit tests.
• e2e tests passed.
• Documentation updated according to the changes.
• Changes were tested in the Kubernetes cluster manually.