Cloudkeel-DD documentation
Cloudkeel-DD is self-hosted configuration-drift detection for teams who run Terraform or Kubernetes on Azure, AWS, or GCP. It compares the state you declared against the state that is actually running, and reports three things: resources that no longer match your code, resources no state declares at all, and resources that break a policy you enabled. It runs inside your own cluster, reads with read-only credentials, and never writes to your cloud.
Where to start
Section titled “Where to start”- Pilot quickstart — install on your own cluster and read your first drift finding.
- Connecting integrations — the desired-vs-actual model, then per-cloud setup with the exact read-only permissions each one needs.
- The drift lifecycle — how a finding is scored, attributed, and resolved.
- Coverage — exactly which resource types are checked field-by-field, and which are tracked as inventory.
Reference
Section titled “Reference”- Act on a drift finding — accept, revert, suppress, or open a remediation PR
- Baselines and suppression — the four ways a finding goes quiet
- Security model — what is read, what is never read, and where credentials live
- Feature inventory — every capability, marked shipped or gap
- Troubleshooting — why an integration has no findings
- Glossary — drift, unmanaged, scope, inventory, spec