Skip to content
This repository has been archived by the owner on Aug 19, 2024. It is now read-only.

Add E2E tests using our examples against real clusters #204

82 changes: 82 additions & 0 deletions .github/workflows/nightly.yaml
Original file line number Diff line number Diff line change
@@ -0,0 +1,82 @@
name: Nightly checks

on:
# workflow_dispatch so that it can be triggered manually if needed
workflow_dispatch:
schedule:
- cron: "34 23 * * *"

jobs:
e2e-tests:
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Here is an example of a similar run (slightly changed to run on this PR, but the idea should be the same): https://github.com/janus-idp/operator/actions/runs/7960039726/job/21728192711?pr=204#step:11:1

runs-on: ubuntu-latest
strategy:
fail-fast: false
matrix:
branch: [ main, 1.1.x ]
name: E2E Tests - ${{ matrix.branch }}
concurrency:
group: ${{ github.workflow }}-${{ matrix.branch }}
cancel-in-progress: true
env:
CONTAINER_ENGINE: podman
steps:
- uses: actions/checkout@v4 # default branch will be checked out by default on scheduled workflows
with:
fetch-depth: 0

- if: ${{ matrix.branch != 'main' }}
name: Checkout ${{ matrix.branch }} branch
run: git switch ${{ matrix.branch }}

# check changes in this commit for regex include and exclude matches; pipe to an env var
- name: Check for changes to build
run: |
# don't fail if nothing returned by grep
set +e
CHANGES="$(git diff --name-only HEAD~1 | \
grep -E "workflows/.+-container-build.yaml|Makefile|bundle/|config/|go.mod|go.sum|.+\.go|docker/|\.dockerignore" | \
grep -v -E ".+_test.go|/.rhdh/")";
echo "Changed files for this commit:"
echo "=============================="
echo "$CHANGES"
echo "=============================="
{
echo 'CHANGES<<EOF'
echo $CHANGES
echo EOF
} >> "$GITHUB_ENV"

- name: Determine built operator image
# run this stage only if there are changes that match the includes and not the excludes
if: ${{ env.CHANGES != '' }}
run: |
SHORT_SHA=$(git rev-parse --short HEAD)
BASE_VERSION=$(grep -E "^VERSION \?=" Makefile | sed -r -e "s/.+= //") # 0.1.0
echo "OPERATOR_IMAGE=quay.io/janus-idp/operator:${BASE_VERSION}-${SHORT_SHA}" >> $GITHUB_ENV

- name: Wait until image exists in registry or timeout is reached
# run this stage only if there are changes that match the includes and not the excludes
if: ${{ env.CHANGES != '' }}
timeout-minutes: 10
run: |
echo "Waiting until operator image is found or timeout expires: ${{ env.OPERATOR_IMAGE }}..."
until ${CONTAINER_ENGINE} image pull "${{ env.OPERATOR_IMAGE }}"; do
sleep 2
echo ...
done
echo "... operator image found: ${{ env.OPERATOR_IMAGE }}."

- name: Start Minikube
# run this stage only if there are changes that match the includes and not the excludes
if: ${{ env.CHANGES != '' }}
uses: medyagh/setup-minikube@606b71970c783154fe49b711486c717f5780f485 # v0.0.15
with:
addons: ingress

- name: Run E2E tests
# run this stage only if there are changes that match the includes and not the excludes
if: ${{ env.CHANGES != '' }}
env:
BACKSTAGE_OPERATOR_TESTS_PLATFORM: minikube
IMG: ${{ env.OPERATOR_IMAGE }}
run: make test-e2e
37 changes: 36 additions & 1 deletion Makefile
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,8 @@ VERSION ?= 0.1.0-dev
# Using docker or podman to build and push images
CONTAINER_ENGINE ?= docker

PKGS := $(shell go list ./... | grep -v /tests)

# CHANNELS define the bundle channels used in the bundle.
# Add a new line here if you would like to change its default config. (E.g CHANNELS = "candidate,fast,stable")
# To re-generate a bundle for other specific channels without changing the standard setup, you can:
Expand Down Expand Up @@ -127,7 +129,7 @@ vet: ## Run go vet against code.
.PHONY: test
test: manifests generate fmt vet envtest ## Run tests. We need LOCALBIN=$(LOCALBIN) to get correct default-config path
mkdir -p $(LOCALBIN)/default-config && cp config/manager/$(CONF_DIR)/* $(LOCALBIN)/default-config
LOCALBIN=$(LOCALBIN) KUBEBUILDER_ASSETS="$(shell $(ENVTEST) use $(ENVTEST_K8S_VERSION) --bin-dir $(LOCALBIN) -p path)" go test ./... -coverprofile cover.out
LOCALBIN=$(LOCALBIN) KUBEBUILDER_ASSETS="$(shell $(ENVTEST) use $(ENVTEST_K8S_VERSION) --bin-dir $(LOCALBIN) -p path)" go test $(PKGS) -coverprofile cover.out

##@ Build

Expand Down Expand Up @@ -227,6 +229,9 @@ GOSEC_VERSION ?= v2.18.2
GOSEC_FMT ?= sarif # for other options, see https://github.com/securego/gosec#output-formats
GOSEC_OUTPUT_FILE ?= gosec.sarif

GINKGO ?= $(LOCALBIN)/ginkgo
GINKGO_VERSION ?= v2.9.5

KUSTOMIZE_INSTALL_SCRIPT ?= "https://raw.githubusercontent.com/kubernetes-sigs/kustomize/master/hack/install_kustomize.sh"
.PHONY: kustomize
kustomize: $(KUSTOMIZE) ## Download kustomize locally if necessary.
Expand Down Expand Up @@ -365,3 +370,33 @@ catalog-update: ## Update catalog source in the default namespace for catalogsou
.PHONY: deploy-openshift
deploy-openshift: release-build release-push catalog-update ## Deploy the operator on openshift cluster

# After this time, Ginkgo will emit progress reports, so we can get visibility into long-running tests.
POLL_PROGRESS_INTERVAL := 120s
TIMEOUT ?= 14400s

GINKGO_FLAGS_ALL = $(GINKGO_TEST_ARGS) --randomize-all --poll-progress-after=$(POLL_PROGRESS_INTERVAL) --poll-progress-interval=$(POLL_PROGRESS_INTERVAL) -timeout $(TIMEOUT) --no-color

# Flags for tests that may be run in parallel
GINKGO_FLAGS=$(GINKGO_FLAGS_ALL) -nodes=$(TEST_EXEC_NODES)
# Flags to run one test per core.
GINKGO_FLAGS_AUTO = $(GINKGO_FLAGS_ALL) -p
ifdef TEST_EXEC_NODES
TEST_EXEC_NODES := $(TEST_EXEC_NODES)
else
TEST_EXEC_NODES := 1
endif

.PHONY: ginkgo
ginkgo: $(GINKGO) ## Download Ginkgo locally if necessary.
$(GINKGO): $(LOCALBIN)
test -s $(LOCALBIN)/ginkgo || GOBIN=$(LOCALBIN) go install github.com/onsi/ginkgo/v2/ginkgo@$(GINKGO_VERSION)

.PHONY: test-e2e
test-e2e: ginkgo ## Run end-to-end tests. See the 'tests/e2e/README.md' file for more details.
$(GINKGO) $(GINKGO_FLAGS) tests/e2e

show-img:
@echo -n $(IMG)

show-container-engine:
@echo -n $(CONTAINER_ENGINE)
2 changes: 1 addition & 1 deletion examples/rhdh-cr-with-app-configs.yaml
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
apiVersion: rhdh.redhat.com/v1alpha1
kind: Backstage
metadata:
name: my-backstage-app-with-app-config
name: bs-app-config
spec:
database:
enableLocalDb: true
Expand Down
135 changes: 135 additions & 0 deletions tests/e2e/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,135 @@
## End-to-end tests

The end-to-end tests use the [Ginkgo framework](https://onsi.github.io/ginkgo/) and allow to test the operator against a real cluster in the following scenarios:
- building and deploying the operator image off of the current code
- using a specific image or a specific downstream build

Deployment of the operator itself can be done by:
- deploying with or without OLM,
- or deploying the downstream bundle in both online and air-gapped scenarios

To run the end-to-end tests, you can use:
```shell
$ make test-e2e
```

### Configuration

The behavior is configurable using the following environment variables:

| Name | Type | Description | Default value | Example |
|------------------------------------------------------------------------------------------------|--------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|---------------------------------------------------|---------------------------------------------------------|
| `BACKSTAGE_OPERATOR_TEST_MODE` | string | The test mode:<br>- if not set, it will call `make deploy`<br>- `olm`: it will call `make deploy-olm`<br>- `rhdh-latest` or `rhdh-next`: it will install the operator using the [`install-rhdh-catalog-source.sh`](../../.rhdh/scripts/install-rhdh-catalog-source.sh) script<br>- `rhdh-airgap`: it will install the operator using the [`prepare-restricted-environment.sh`](../../.rhdh/scripts/prepare-restricted-environment.sh) script. | | `rhdh-latest` |
| `IMG` (or any variables from the Makefile that are used by `make deploy` or `make deploy-olm`) | string | The image to use. Relevant if `BACKSTAGE_OPERATOR_TEST_MODE` is not set or set to `olm` | `VERSION` defined in [`Makefile`](../../Makefile) | `quay.io/janus-idp/operator:0.0.1-latest` |
| `BACKSTAGE_OPERATOR_TESTS_BUILD_IMAGES` | bool | If set to `true`, it will build the operator image with `make image-build`.<br>Relevant if `BACKSTAGE_OPERATOR_TEST_MODE` is not set or set to `olm`. | | `false` |
| `BACKSTAGE_OPERATOR_TESTS_PUSH_IMAGES` | bool | If set to `true`, it will push the operator image with `make image-push`.<br>Relevant if `BACKSTAGE_OPERATOR_TEST_MODE` is not set or set to `olm`. | | `false` |
| `BACKSTAGE_OPERATOR_TESTS_PLATFORM` | string | The platform type, to directly load the operator image if supported instead of pushing it.<br>Relevant if `BACKSTAGE_OPERATOR_TEST_MODE` is not set or set to `olm`.br>Supported values: [`kind`](#building-and-testing-local-changes-on-kind), [`k3d`](#building-and-testing-local-changes-on-k3d), [`minikube`](#building-and-testing-local-changes-on-minikube) | | `kind` |
| `BACKSTAGE_OPERATOR_TESTS_KIND_CLUSTER` | string | Name of the local KinD cluster to use. Relevant only if `BACKSTAGE_OPERATOR_TESTS_PLATFORM` is `kind`. | `kind` | `kind-local-k8s-cluster` |
| `BACKSTAGE_OPERATOR_TESTS_K3D_CLUSTER` | string | Name of the local k3d cluster to use. Relevant only if `BACKSTAGE_OPERATOR_TESTS_PLATFORM` is `k3d`. | `k3s-default` | `k3d-local-k8s-cluster` |
| `BACKSTAGE_OPERATOR_TESTS_AIRGAP_INDEX_IMAGE` | string | Index image to use in the airgap scenario.<br>Relevant if `BACKSTAGE_OPERATOR_TEST_MODE` is `rhdh-airgap`. | `quay.io/rhdh/iib:latest-v4.14-x86_64` | `registry.redhat.io/redhat/redhat-operator-index:v4.14` |
| `BACKSTAGE_OPERATOR_TESTS_AIRGAP_OPERATOR_VERSION` | string | Operator version to use in the airgap scenario.<br>Relevant if `BACKSTAGE_OPERATOR_TEST_MODE` is `rhdh-airgap`. | `v1.1.0` | `v1.1.0` |
| `BACKSTAGE_OPERATOR_TESTS_AIRGAP_MIRROR_REGISTRY` | string | Existing mirror registry to use in the airgap scenario.<br>Relevant if `BACKSTAGE_OPERATOR_TEST_MODE` is `rhdh-airgap`<br>. | | `my-registry.example.com` |

### Examples

#### Testing the operator available for the VERSION (default)

In this scenario, you want to run the E2E test suite against the operator image corresponding to the `VERSION` declared in the project [`Makefile`](../../Makefile), which should be publicly available at `quay.io/janus-idp/operator:<VERSION>`.

This is the default behavior.

This should work on any Kubernetes or OpenShift cluster:

```shell
$ make test-e2e
```

#### Testing a specific image (e.g. PR image)

In this scenario, you want to run the E2E test suite against an existing operator image.

This should work on any Kubernetes or OpenShift cluster:

```shell
# if the tag is already published and available at the default location: quay.io/janus-idp/operator
$ make test-e2e VERSION=0.2.0-3d1c1e0

# or you can override the full image repo name
$ make test-e2e IMG=my.registry.example.com/operator:0.2.0-3d1c1e0
```

Note that `VERSION` and `IMG` override the respective variables declared in the project [`Makefile`](../../Makefile).

#### Building and testing local changes on supported local clusters

In this scenario, you are iterating locally, and want to run the E2E test suite against your local changes. You are already using a local cluster like [`kind`](https://kind.sigs.k8s.io/), [`k3d`](https://k3d.io/) or [`minikube`](https://minikube.sigs.k8s.io/docs/), which provide the ability to import images into the cluster nodes.

To do so, you can:
1. set `BACKSTAGE_OPERATOR_TESTS_BUILD_IMAGES` to `true`, which will result in building the operator image from the local changes,
2. and set `BACKSTAGE_OPERATOR_TESTS_PLATFORM` to a supported local cluster, which will result in loading the image built directly in that cluster (without having to push to a separate registry).

##### `kind`

```shell
$ kind create cluster
$ make test-e2e \
BACKSTAGE_OPERATOR_TESTS_BUILD_IMAGES=true \
BACKSTAGE_OPERATOR_TESTS_PLATFORM=kind
```

##### `k3d`

```shell
$ k3d cluster create
$ make test-e2e \
BACKSTAGE_OPERATOR_TESTS_BUILD_IMAGES=true \
BACKSTAGE_OPERATOR_TESTS_PLATFORM=k3d
```

##### `minikube`

```shell
$ minikube start
$ make test-e2e \
BACKSTAGE_OPERATOR_TESTS_BUILD_IMAGES=true \
BACKSTAGE_OPERATOR_TESTS_PLATFORM=minikube
```

#### Testing a specific version using OLM

In this scenario, you want to leverage the [Operator Lifecycle Manager (OLM)](https://olm.operatorframework.io/) to deploy the Operator.

This requires OLM to be installed in the cluster.

```shell
$ make test-e2e BACKSTAGE_OPERATOR_TEST_MODE=olm
```

#### Testing a downstream build of Red Hat Developer Hub (RHDH)

In this scenario, you want to run the E2E tests against a downstream build of RHDH.

This works only against OpenShift clusters. If testing a CI build, please follow the instructions in [Installing CI builds of Red Hat Developer Hub](../../.rhdh/docs/installing-ci-builds.adoc) to add your Quay token to the cluster.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If it needs that this Openshift cluster be configured as a current context of your kubeconfig, could you please describe this a bit... It is not really clear how 'make test-e2e BACKSTAGE_OPERATOR_TEST_MODE=rhdh-latest' knows where this Openshift is :)

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

But this is out of scope of this document, no? Anyway, I've added a note and link to help clarify this. See ff4daa1 (#204)

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I am just trying to make it clear for myself... not sure about the doc scoping :)
I think description counle be as smple as: "Make sure your current context points to..... calling 'kubectl config current-context'"

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Okay. See 2b5b821 (#204)

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If you are talking about "To run the end-to-end tests, make sure you have an active connection to a cluster in your current [kubeconfig]"
I am just not sure about terminology like "active connection" in this context :) , to me "current context " sounds more familiar in this context.
But if you think it is clear enough - go ahead, I am not blocking it.

Copy link
Member Author

@rm3l rm3l Feb 22, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If you are talking about "To run the end-to-end tests, make sure you have an active connection to a cluster in your current [kubeconfig]"
I am just not sure about terminology like "active connection" in this context :) , to me "current context " sounds more familiar in this context.

Yeah, I borrowed the words "active connection" from the OpenShift doc. Then added a note on how to check the current context, as you suggested ;-)

But if you think it is clear enough - go ahead, I am not blocking it.

Okay, please approve the PR then if it is good for you ^^. Thanks.


```shell
# latest
$ make test-e2e BACKSTAGE_OPERATOR_TEST_MODE=rhdh-latest

# or next
$ make test-e2e BACKSTAGE_OPERATOR_TEST_MODE=rhdh-next
```

#### Airgap testing of RHDH

In this scenario, you want to run the E2E tests against an OpenShift cluster running in a restricted network. For this, the command below will make sure to prepare it by copying all the necessary images to a mirror registry, then deploy the operator.
Please make sure to read the prerequisites in [Installing Red Hat Developer Hub (RHDH) in restricted environments](../../.rhdh/docs/airgap.adoc).

```shell
# if you want to have a mirror registry to be created for you as part of the airgap environment setup
$ make test-e2e BACKSTAGE_OPERATOR_TEST_MODE=rhdh-airgap

# or if you already have a mirror registry available and reachable from within your cluster
$ make test-e2e \
BACKSTAGE_OPERATOR_TEST_MODE=rhdh-airgap \
BACKSTAGE_OPERATOR_TESTS_AIRGAP_MIRROR_REGISTRY=my-registry.example.com
```
Loading
Loading