fix(main): drop forced --allow-untagged-cloud requirement to silence deprecation warning

[arbianshkodra]

What

Removes the fail-fast in cloudInitializer (main.go) that forced operators to pass --allow-untagged-cloud. HasClusterID() is left unchanged (still returns false).

Why

The kubernetes/cloud-provider library deprecated --allow-untagged-cloud and emits a startup warning whenever the flag is set:

Flag --allow-untagged-cloud has been deprecated, This flag is deprecated and will be removed in a future release. A cluster-id will be required on cloud instances.

For Hetzner, no ClusterID is ever consumed by the cloud provider — confirmed in #1119:

We don't use cluster-ids anywhere in the code. Neither has allow-untagged-cloud any relevance for us. ... You can safely ignore this message from the logs. We may want to update our code here and don't emit any messages.
— @lukasmetzner

Because HasClusterID() returns false, cloudInitializer would klog.Fatalf on startup unless --allow-untagged-cloud was set. Operators were therefore forced to pass the flag — and passing the (deprecated) flag is exactly what triggered the upstream warning. The flag was functionally a no-op for Hetzner; its mere presence caused the warning.

By removing the fail-fast requirement, operators can omit the flag entirely and the deprecation warning disappears at its source (the upstream library only emits it when the flag is set).

Why not change HasClusterID()?

The initial approach changed HasClusterID() to return true. Per review feedback from @lukasmetzner, that value is part of the upstream cloud-provider interface and changing it could cause unexpected behavior if consumed upstream. This revision keeps the interface untouched and instead drops the locally-owned fail-fast in main.go.

Behavior change

Before	After
Operator must set --allow-untagged-cloud or HCCM crashes on startup	Operator can omit the flag
Deprecation warning fires when the flag is set (every install today)	No warning fires when the flag is omitted
HasClusterID() returns false	HasClusterID() returns false (unchanged)

Operators who keep --allow-untagged-cloud set will continue to see the deprecation warning (the upstream library still notices the flag) — but they no longer need the flag.

Testing

• go build ./... passes
• go test ./hcloud/ -run TestCloud passes

Closes #1119

[lukasmetzner]

Hey @arbianshkodra,

I'd prefer not to change the value returned by HasClusterID(), since it's part of the cloud provider interface we implement and could cause unexpected behavior if used upstream.

That said, I'm fine with removing the startup warning, as I don't see much upstream movement toward using cluster IDs.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 67.17%. Comparing base (547798d) to head (43b4e29).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1244      +/-   ##
==========================================
- Coverage   71.01%   67.17%   -3.84%     
==========================================
  Files          24       24              
  Lines        2677     2672       -5     
==========================================
- Hits         1901     1795     -106     
- Misses        604      706     +102     
+ Partials      172      171       -1     

Flag	Coverage Δ
e2e	?
unit	67.17% <ø> (+0.12%)	⬆️

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

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[lukasmetzner]

The e2e tests failing is expected here, due to our permissions.

[arbianshkodra]

@lukasmetzner, understood. Should I update this?

[lukasmetzner]

Yes, feel free update this PR.

[arbianshkodra]

Done. Kept HasClusterID() as-is and dropped the forced --allow-untagged-cloud requirement in main.go instead. Warning's gone.