Existing Apps

If you already run Helm, Argo, Flux, or plain Kubernetes YAML, the first problem is ownership, not migration. Start read-only. First understand what exists. Then decide what ConfigHub will manage.

Start Read-Only

Existing systems often have history: old chart versions, local patches, hand-created Secrets, controller-generated fields, or cluster-specific assumptions. ConfigHub makes those facts visible before it tries to manage them.

The first safe outcome is an inventory and comparison, not a changed live deployment.

Starting pointFirst routeBoundary
Argo or Flux appStart by reading the current source, rendered objects, target namespace, health, and sync state.Do not change delivery yet. Compare what exists with a catalog or recipe path first.
Rendered YAMLImport or inspect the object set as desired state data.Check object identity, labels, namespaces, Secrets, CRDs, and hooks before trying to manage it.
Live clusterUse observation first: what is running, who owns it, what changed, and what target facts are required?Treat live state as evidence, not automatically as desired state.
Helm releaseKeep the chart, version, values, and release name as the starting facts.Then decide whether the first ConfigHub base matches that release exactly or intentionally differs.

What To Check First

CheckWhy it matters
IdentityWhich chart, app, namespace, target, and owner does this belong to?
Object setWhich Deployments, Services, CRDs, RBAC, Secrets, ConfigMaps, and policies exist?
DifferencesWhat does the catalog recipe render for the same chart and values?
PrerequisitesWhich Secrets, storage classes, cloud identities, CRDs, and controllers must already exist?
Control pointWhat is the safest first managed change: observe only, create a base, create a derived variant, or promote a patch?

Where This Leads

Match the current app

Create or select a base that renders the same object set as the existing Helm release.

Existing app guide

Create a managed variant

Once the base is trusted, use a derived variant for environment, region, customer, or target-specific refinements.

Variants

Move into operations

After upload, use scans, approvals, delivery, observations, upgrades, and rollback records.

Ops