An operator for managing a set of wasmCloud hosts running on Kubernetes and manage wasmCloud applications using wadm. The goal is to easily be able to run WasmCloud hosts on a Kubernetes cluster.
The WasmCloudHostConfig CRD describes the desired state of a set of wasmCloud hosts connected to the same lattice.
apiVersion: k8s.wasmcloud.dev/v1alpha1
kind: WasmCloudHostConfig
metadata:
name: my-wasmcloud-cluster
spec:
# The number of wasmCloud host pods to run
hostReplicas: 2
# The lattice to connect the hosts to
lattice: default
# Additional labels to apply to the host other than the defaults set in the operator
hostLabels:
some-label: value
# The address to connect to nats
natsAddress: nats://nats.default.svc.cluster.local
# Which wasmCloud version to use
version: 1.0.4
# Enable the following to run the wasmCloud hosts as a DaemonSet
#daemonset: true
# The name of the image pull secret to use with wasmCloud hosts so that they
# can authenticate to a private registry to pull components.
# registryCredentialsSecret: my-registry-secret
The CRD requires a Kubernetes Secret with the following keys:
apiVersion: v1
kind: Secret
metadata:
name: my-wasmcloud-cluster
#data:
# Only required if using a NATS creds file
# nats.creds: <creds file>
The operator will fail to provision the wasmCloud Deployment if any of these secrets are missing!
If you would like to customize the registry or image that gets used to provision the wasmCloud hosts and the NATS leaf that runs alongside them, you can specify the following options in the above WasmCloudHostConfig
CRD.
For wasmCloud Host, use the image
field:
apiVersion: k8s.wasmcloud.dev/v1alpha1
kind: WasmCloudHostConfig
metadata:
name: my-wasmcloud-cluster
spec:
# other config options omitted
image: registry.example.com/wasmcloud:1.0.2
For the NATS leaf, use the natsImageLeaf
field:
apiVersion: k8s.wasmcloud.dev/v1alpha1
kind: WasmCloudHostConfig
metadata:
name: my-wasmcloud-cluster
spec:
# other config options omitted
natsLeafImage: registry.example.com/nats:2.10.16
You can also specify an image pull secret to use use with the wasmCloud hosts
so that they can pull components from a private registry. This secret needs to
be in the same namespace as the WasmCloudHostConfig CRD and must be a
kubernetes.io/dockerconfigjson
type secret. See the Kubernetes
documentation
for more information on how to provision that secret.
Once it is created, you can reference it in the WasmCloudHostConfig CRD by
setting the registryCredentialsSecret
field to the name of the secret.
A wasmCloud cluster requires a few things to run:
- A NATS cluster with Jetstream enabled
- WADM connected to the NATS cluster in order to support applications
If you are running locally, you can use the following commands to start a NATS cluster and WADM in your Kubernetes cluster.
Use the upstream NATS Helm chart to start a cluster with the following values.yaml file:
config:
cluster:
enabled: true
replicas: 3
leafnodes:
enabled: true
jetstream:
enabled: true
fileStore:
pvc:
size: 10Gi
merge:
domain: default
helm repo add nats https://nats-io.github.io/k8s/helm/charts/
helm upgrade --install -f values.yaml nats nats/nats
You can run Wadm in your Kubernetes cluster using our Helm chart. For a minimal deployment using the
NATS server deployed above, all you need in your values.yaml
file is:
wadm:
config:
nats:
server: "nats.default.svc.cluster.local:4222"
You can deploy Wadm using your values file and Helm:
helm install wadm -f wadm-values.yaml --version 0.2.0 oci://ghcr.io/wasmcloud/charts/wadm
kubectl kustomize deploy/base | kubectl apply -f -
The operator automatically creates Kubernetes Services for wasmCloud
applications. Right now this is limited only to applications that deploy the
wasmCloud httpserver component using a daemonscaler
, but additional support
for spreadscalers
will be added in the future.
If you specify host label selectors on the daemonscaler
then the operator
will honor those labels and will only create a service for the pods that match
those label selectors.
Argo CD provides a way to define a custom health check that it then runs against a given resource to determine whether or not the resource is in healthy state.
For this purpose, we specifically expose a status.phase
field, which exposes
the underlying status information from wadm.
With the following ConfigMap, a custom health check can be added to an existing Argo CD installation for tracking the health of wadm applications.
---
apiVersion: v1
kind: ConfigMap
metadata:
name: argocd-cm
namespace: argocd
labels:
app.kubernetes.io/name: argocd-cm
app.kubernetes.io/part-of: argocd
data:
resource.customizations: |
core.oam.dev/Application:
health.lua: |
hs = {}
hs.status = "Progressing"
hs.message = "Reconciling application state"
if obj.status ~= nil and obj.status.phase ~= nil then
if obj.status.phase == "Deployed" then
hs.status = "Healthy"
hs.message = "Application is ready"
end
if obj.status.phase == "Reconciling" then
hs.status = "Progressing"
hs.message = "Application has been deployed"
end
if obj.status.phase == "Failed" then
hs.status = "Degraded"
hs.message = "Application failed to deploy"
end
if obj.status.phase == "Undeployed" then
hs.status = "Suspended"
hs.message = "Application is undeployed"
end
end
return hs
- Make sure you have a Kubernetes cluster running locally. Some good options include Kind or Docker Desktop.
RUST_LOG=info cargo run
This repo stores the types for any CRDs used by the operator in a separate
crate (wasmcloud-operator-types
) so that they can be reused in other projects.