> For the complete documentation index, see [llms.txt](https://developer.harness.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://developer.harness.io/continuous-delivery/troubleshooting-and-resources/armory/general/cannot-add-custom-kind-to-delete-manifest-stage.md).

# Cannot add custom Kind to Delete Manifest Stage

### Issue <a href="#issue" id="issue"></a>

When adding the Delete Manifest Stage, Spinnaker will require a `Kind` field. For the Deploy Manifest Stage, a custom `Kind` can be added. However, Spinnaker is not able to get the Kubernetes custom `Kinds` for the `Delete Manifest Stage`. When trying to add a custom `Kind` to the `Delete Manifest Stage`, the user may see an error such as: Exception ( Monitor Delete ) Failed to delete none/ from cluster: error: the server doesn't have a resource type "none"

### Cause <a href="#cause" id="cause"></a>

The `Delete Manifest stage` works differently within Spinnaker than the Deploy Manifest stage. The `KubernetesDeleteManifestOperatation` that is used during the `Delete Manifest stage` uses the `kind`'s handler to do the deletion. All unregistered `kinds` use the `KubernetesUnregisteredCustomResourceHandler`, which always sets the `kind` to `NONE.` This isn't a problem for deploy operations as it doesn't need to do a lookup to find an existing resource, it is just adding the new resource to the cluster. For the delete operations, the current `kind` needs to be sent as part of the delete operation, but the handler only knows the `kind` as `NONE`.