woozymasta/kube-dump repository overview

⁠kube-dump

Git project: WoozyMasta/kube-dump⁠

License: MIT

Documentation site⁠ · Run in Kubernetes⁠ · Project documentation⁠

⁠kube-dump

kube-dump is a utility for preserving the actual state of a Kubernetes environment when it needs to be captured.

In a test namespace, developer sandbox, temporary staging environment, or demonstration environment, state is often changed manually. Someone updates a Deployment, creates a ConfigMap, tries a new operator, writes data to a PVC, or runs a temporary image. That state may be needed again later, even if the environment has already changed or been deleted.

These environments do not always have GitOps, a separate backup system, or a strict lifecycle. That is fine, but important state can still be left without a portable copy.

kube-dump preserves this state in a readable format that can be stored locally, in Git, or in S3-compatible storage and used for recovery.

⁠Installation

Ready-to-use files for the latest stable release:

All published versions are available on GitHub Releases⁠.

With Go 1.27 or newer, install the current version with:

go install github.com/woozymasta/kube-dump/v2/cmd/[email protected]

For running inside Kubernetes, use the ready-made Kustomize components⁠.

The container image is published to:

ghcr.io/woozymasta/kube-dump:2.0.0-rc.1
docker.io/woozymasta/kube-dump:2.0.0-rc.1

For CI, debugging, and scripting, use the image with a shell and additional utilities:

ghcr.io/woozymasta/kube-dump:2.0.0-rc.1-debug
docker.io/woozymasta/kube-dump:2.0.0-rc.1-debug

More installation options are described in the documentation⁠.

⁠What it preserves

kube-dump preserves three independent parts of Kubernetes state:

• Kubernetes resources - YAML manifests for storing in Git, S3, or an archive;
• PVC data - file archives that can be read from a CSI snapshot;
• container images - a standard OCI Image Layout, including temporary images and merge request builds.

Each layer can be saved separately or together with the others.

⁠How to use it

Each data type has its own command. For example, save resources locally:

kube-dump resource save dir ./backup

Or send them to S3-compatible storage:

kube-dump resource save s3 --s3-uri s3://backup/my-cluster/

Save PVC data and container images with the pvc and image commands.

Capture the state before maintenance or an experiment, before cleaning a temporary registry, or before deleting a test environment.

⁠What it does not replace

kube-dump does not replace Argo CD, Flux, Velero, database operators, or an application's own backup mechanisms.

If an application can create a correct logical database backup, continue to use that mechanism. If infrastructure is fully declared in Git, Git remains the source of desired state.

kube-dump captures what actually exists in the cluster at a given moment in a readable and portable form, so it can be used as a basis for recovery.

⁠Profiles and data protection

Profiles define how resources, PVCs, and images are selected and processed. For Kubernetes manifests, they also define cleanup rules and fields to encrypt.

Built-in profiles:

• backup⁠ - backs up configuration, access control, and common operator objects while removing server-generated state;
• export⁠ - creates clean, portable manifests without runtime metadata and empty fields;
• raw⁠ - saves all discovered objects without cleanup or filtering.

Sensitive YAML values can be encrypted with age or deterministic AES-256-SIV. Resource and PVC archives can additionally be protected with age.

⁠More information

This README is not the complete documentation. Installation, commands, profiles, encryption, PVCs, and images are covered in the documentation⁠.