Skip to content

chart.md

Latest commit

 

History

History
118 lines (84 loc) · 3.6 KB

chart.md

File metadata and controls

118 lines (84 loc) · 3.6 KB
Raw

Image Factory Helm Chart

If you want to deploy Image Factory on a Kubernetes cluster, you can use the provided Helm chart located in the deploy/helm/image-factory directory. It is also distributed via the OCI registry.

E2E Tests

Image Factory uses Chainsaw integration tests for the Image Factory Helm chart.

Prerequisites

  1. Install Chainsaw:

    make chainsaw-install

    This downloads the chainsaw CLI into _out/.

  2. Helm: Helm v4+ installed

Running Tests

Before running any tests, prepare your development cluster. You can do this by running:

make k8s-up

To tear down the cluster after testing, run:

make k8s-down

Run all tests

export KUBECONFIG="$(pwd)/_out/kubeconfig"
cd deploy/helm/e2e
$(pwd)/../../../_out/chainsaw test

Or via the Makefile target:

make chart-e2e-chainsaw

Run a single test

cd deploy/helm/e2e
$(pwd)/../../../_out/chainsaw test tests/01-image-factory

Run with custom timeouts

chainsaw test --exec-timeout 600s --assert-timeout 600s

Configuration

Test Suite Settings (deploy/helm/e2e/.chainsaw.yaml)

  • timeouts.apply: 60s - per-apply operation timeout
  • timeouts.assert: 600s - per-assert operation timeout
  • timeouts.cleanup: 300s - cleanup timeout
  • timeouts.exec: 600s - script/command timeout (covers helm install --wait)
  • execution.parallel: 1 - tests run sequentially
  • execution.failFast: true - stop on first failure
  • templating.enabled: true - enables ($namespace) jmespath expressions in resources

Each test runs in its own ephemeral namespace (chainsaw-<random>) which Chainsaw creates and deletes automatically.

Layout

deploy/helm/e2e/
├── .chainsaw.yaml          # root configuration
├── _crds/                  # cluster-scoped CRDs applied as test setup
├── _manifests/             # cluster-scoped manifests (local-path-storage, ...)
├── _lib/                   # shared StepTemplates referenced via `use:` in tests
├── testdata/               # signing keys, htpasswd, cosign keys
└── tests/
    ├── 01-image-factory/   # basic install/upgrade/uninstall
    ├── 02-upstream/        # internal registry, schematic from upstream
    ├── 03-airgapped/       # mirror + cosign Talos images, airgapped install
    └── 04-enterprise/      # enterprise build (auth, grypeDB PVC)

Debugging

export KUBECONFIG=./_out/kubeconfig
kubectl get ns | grep chainsaw-       # find ephemeral namespace
kubectl get all -n chainsaw-<suffix>
kubectl logs -n chainsaw-<suffix> deployment/image-factory

Failing steps automatically dump describes and pod logs via Chainsaw catch: blocks defined in _lib/*.step.yaml and per-test try.catch sections.

To keep resources around for inspection after a failure, pass --skip-delete:

chainsaw test --skip-delete --pause-on-failure

Notes

  • Prometheus Operator CRDs are pre-applied from _crds/ by the 01-image-factory test (required for ServiceMonitor assertions).
  • local-path-storage is pre-applied from _manifests/ by the 04-enterprise test (required for the grypeDB PVC).
  • Tests use ephemeral namespaces; service references in tests resolve the namespace via the ($namespace) jmespath binding.
  • Helm upgrades use --reuse-values to maintain state across steps.
  • The ECDSA signing key is provided via a pre-generated key file in testdata/.
  • Cosign keys for the airgapped test are in testdata/cosign.key and testdata/cosign.pub.