The intention of this guide is to give you more detailed information on various
elements of the
HelmRelease Custom Resource. You can follow it in one take,
use it as a more explanatory reference, or a combination of both.
It assumes you have the Helm Operator already installed in your cluster. If you have not done this yet, follow the installation instructions from the quickstart
The guide tries not to presume a specific enabled Helm version but for the sake of simplicity it was written while making use of Helm 3.
A minimal HelmRelease¶
To get started, we will first create the most minimal
We will use (parts) of this minimal
HelmRelease as an example throughout the
rest of this guide.
# podinfo.yaml --- apiVersion: helm.fluxcd.io/v1 kind: HelmRelease metadata: name: podinfo namespace: default spec: chart: repository: https://stefanprodan.github.io/podinfo name: podinfo version: 3.2.0
spec.chart object is the only mandatory property of the
and defines the Helm chart that should be installed by the Helm Operator. This
HelmRelease will manage
a tiny web application made with Go that showcases best practices of running microservices in Kubernetes
from a Helm repository chart source.
Applying this to the cluster and making the Helm Operator do the release is
equivalent to equals to a human running the following
$ helm repo add podinfo https://stefanprodan.github.io/podinfo $ helm upgrade -i default-podinfo podinfo/podinfo --version '3.2.0'
A couple of differences between the
HelmRelease resource and listed
commands should stand out:
HelmReleasedoes not make use of a repository alias, instead the absolute URL of the Helm repository is used.
This is to ensure the
HelmReleaseis able to stand on its own. If we used names in the spec, which were resolved to URLs elsewhere (e.g., in a
repositories.yamlsupplied to the operator), it would be possible to change the meaning of a
HelmReleasewithout altering it. This is undesirable because it makes it hard to specify exactly what you want, in the one place; or to read exactly what is being specified, in the one place.
The reasoning behind this is the same as the explanation given above.
- The default release name used by the Helm Operator is composed from the
This to ensure it does not collide with other
HelmReleaseresources that may have the same
.metadata.namebut different namespace definitions.
Having absorbed this information, you can now go ahead and apply the resource
$ kubectl apply -f podinfo.yaml
Take a look at the logs of the Helm Operator deployment and you will see the Helm installation was performed almost instantly after applying the resource:
$ kubectl logs deploy/flux-helm-operator ... ts=2020-01-01T12:00:00.000000000Z caller=release.go:335 component=release release=default-podinfo targetNamespace=default resource=default:helmrelease/podinfo helmVersion=v3 info="no existing release" action=install ts=2020-01-01T12:00:00.000000000Z caller=helm.go:69 component=helm version=v3 info="creating 2 resource(s)" targetNamespace=default release=default-podinfo ts=2020-01-01T12:00:00.000000000Z caller=release.go:266 component=release release=default-podinfo targetNamespace=default resource=default:helmrelease/podinfo helmVersion=v3 info="Helm release sync succeeded" revision=3.2.0 $ kubectl get helmrelease NAME RELEASE STATUS MESSAGE AGE podinfo default-podinfo deployed Helm release sync succeeded 10s
As expected, the release will now also show up for e.g.
$ helm list NAME NAMESPACE REVISION UPDATED STATUS CHART APP VERSION default-podinfo default 1 2020-01-01 12:00:00.000000000 +0000 UTC deployed podinfo-3.2.0 3.2.0
Congratulations! You made your first Helm release using a