chore(cloudprem): bump to 0.4.1 with multi-cloud docs + kubeVersion enforcement

[Mallets]

Summary

Doc-audit follow-up surfaced by an audit of the public BYOC Logs docs against the chart and the pomsky storage layer. Bundles four changes to the cloudprem chart so the README + Chart.yaml match the chart's actual multi-cloud support and the public docs.

Audit evidence: https://github.com/DataDog/claude-marketplace/pull/2216#issuecomment-4543572562

Changes

1. kubeVersion: ">= 1.25.0-0" added to Chart.yaml. Today the K8s 1.25+ requirement is documented in two places (README Prereqs and the public install page) but not enforced — helm install accepts older clusters and customers hit obscure runtime failures. The -0 suffix accepts cloud-vendor variants like 1.25.0-eks-abc.

2. README Prerequisites section rewritten from AWS-only to multi-cloud:

   • The chart's values.yaml has had a full azure: block; the public install docs cover EKS / GKE / AKS / self-managed K8s; quickwit-storage registers S3 + Azure + GCS + S3-compat factories.
   • The README listing "AWS account / AWS Load Balancer Controller / S3 bucket" as bullet prereqs was out of sync with all of those.
   • New Prereqs cover supported clouds, both supported ingress controllers (ALB + NGINX), the per-cloud PostgreSQL options, and the full object-store list.

3. image.tag default in the values table — was documented as devel, actually defaults to the chart's appVersion when unset (currently v0.1.26). Customers reading the README expected one tag and got another.

4. azure.* rows added to the values table so the Azure configuration surface is discoverable from the README. (No gcp.* rows because GCP uses Workload Identity, wired via serviceAccount.extraAnnotations — no chart-key gap there.)

Both README.md and README.md.gotmpl are edited so the change survives whichever path future regeneration takes.

Test plan

• helm lint charts/cloudprem passes locally.
• helm install against a K8s 1.24 cluster now fails with chart requires kubeVersion ">= 1.25.0-0".
• helm install against a K8s 1.25+ cluster still installs cleanly with no image.tag override.
• Visual review of the regenerated README.md (e.g. via artifacthub preview) matches the values-table additions.
• CHANGELOG entry under 0.4.1 reads cleanly to the chart-team reviewer.

🤖 Generated with Claude Code

[datadog-official]

✨ Fix all issues with BitsAI

⚠️ Warnings

🚦 2 Pipeline jobs failed

Lint and Test Charts | pr-validated     

🔧 Fix in code (Fix with Cursor).

kubeconform-chart installation failed with exit code 1.

Lint and Test Charts | kubeconform-chart (v1.18.20)     

🛟 This job is unlikely to succeed on retry. Please review your pipeline configuration.

Chart requires kubeVersion: >= 1.25.0-0 which is incompatible with Kubernetes v1.18.20.

Useful? React with 👍 / 👎

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: 7091313 | Docs | Datadog PR Page | Give us feedback!}

[Mallets]

CI status — initial run

Pushed 7091313 to fix the only legitimate failure (lint-docs); the rest of the red are expected.

Check	Status	Note
lint-docs	🟢 should pass after 7091313	helm-docs auto-expands {{ template "chart.requirementsSection" . }} (already present in README.md.gotmpl:14) into a ## Requirements block now that Chart.yaml declares kubeVersion. The committed README.md was missing the rendered output. The .gotmpl itself was already correct.
kubeconform-chart (v1.16.4 → v1.24.17)	🔴 expected — this is the feature working	Logs show Error: chart requires kubeVersion: >= 1.25.0-0 which is incompatible with Kubernetes v<X>. That is exactly what the kubeVersion constraint is meant to do at install time. The CI matrix lints every chart against every K8s version regardless of declared kubeVersion.
kubeconform-chart (v1.25.16 → v1.31.1)	🟢	All supported versions pass.
All other checks	🟢	Lint, version bump, mergegate, etc. — all green.

On the pre-1.25 kubeconform failures

I deliberately did not soften the kubeVersion constraint to make these pass, because that would defeat the whole point of the change (install-time rejection rather than runtime breakage on EOL K8s versions). Two follow-up options for the chart team to decide:

1. Per-chart matrix exclusion — update the workflow to skip kubeconform matrix entries for a chart when those versions fall outside the chart's declared kubeVersion. Cleanest long-term, but a workflow-wide change.
2. Drop pre-1.25 versions from the matrix repo-wide — K8s 1.16 was EOL in 2020-09, 1.18 in 2021-04, 1.22 in 2022-12, 1.24 in 2024-07. None are supported by any current cloud-managed K8s service. Reduces matrix run time too.

Happy to open a follow-up PR for either, but it touches the workflow file (.github/workflows/lint-test.yaml), which is out of scope for this audit follow-up.