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 cluster issuers to use for each host
issuers:
# This needs to be the public key of a cluster seed generated by
# `wash keys gen cluster`
- CDK...
# The lattice to connect the hosts to
lattice: 83a5b52e-17cf-4080-bac8-f844099f142e
# Additional labels to apply to the host other than the defaults set in the operator
hostLabels:
some-label: value
# Which wasmCloud version to use
version: 0.82.0
# The name of a secret in the same namespace that provides the required secrets.
secretName: cluster-secrets
# 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:
# You can generate this with wash:
# `wash keys gen cluster`
WASMCLOUD_CLUSTER_SEED: <seed>
# Only required if using a NATS creds file
# nats.creds: <base64 encoded creds file>
The operator will fail to provision the wasmCloud Deployment if any of these secrets are missing!
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 upgrade --install -f values.yaml nats-cluster 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-cluster.default.svc.cluster.local:4222"
You can deploy Wadm using your values file and Helm:
helm install wadm -f values.yaml --version 0.2.0 oci://ghcr.io/wasmcloud/charts/wadm
kubectl kustomize build deploy/local | 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.
27 Sep 8, 2022
3 Feb 26, 2024
10.4k Jan 4, 2023
2 Oct 16, 2022
1 Apr 21, 2022
2 Jul 23, 2022
20 Jun 5, 2023
2 Nov 20, 2022
5 Jul 26, 2022
1 Feb 1, 2022
5 Jan 25, 2023
42 May 23, 2023
25 Apr 16, 2023
18 Sep 23, 2024
2 Jul 30, 2022
52 Apr 9, 2023
36 Jan 28, 2023
8 Feb 10, 2022
156 Nov 21, 2024