Add-ons are Addon objects (harvesterhci.io/v1beta1) wrapping Helm charts, enabled/disabled from the UI (Advanced → Add-ons).
| Add-on | Purpose |
|---|---|
rancher-monitoring |
Prometheus/Grafana/Alertmanager for the Harvester cluster |
rancher-logging |
Logging operator; ship host/k8s/audit/event logs to outputs |
vm-import-controller |
Import VMs from VMware vSphere and OpenStack |
pcidevices-controller |
PCI/USB/SR-IOV/vGPU passthrough |
nvidia-driver-toolkit |
NVIDIA vGPU host drivers |
harvester-seeder |
Out-of-band management (IPMI/Redfish), inventory, hardware events |
harvester-csi-driver-lvm |
Local LVM storage |
descheduler |
Rebalance VMs across nodes |
rancher-vcluster |
Run a full Rancher Manager inside a vcluster on Harvester (experimental; handy for small sites) |
kubeovn-operator / networking add-ons |
Overlay/VPC networking (version-dependent naming) |
Configure add-on values the same way as Helm values:
apiVersion: harvesterhci.io/v1beta1
kind: Addon
metadata:
name: rancher-logging
namespace: cattle-logging-system
spec:
enabled: true
repo: http://harvester-cluster-repo.cattle-system.svc/charts
chart: rancher-logging
version: "<bundled version>"
valuesContent: |
fluentbit:
resources:
limits:
memory: 256Mi
Logging example — ship Harvester logs to Loki or Elasticsearch:
apiVersion: logging.banzaicloud.io/v1beta1
kind: ClusterOutput
metadata:
name: loki
namespace: cattle-logging-system
spec:
loki:
url: http://loki.observability.lab.example.net:3100
extra_labels:
cluster: harvester-prod
configure_kubernetes_labels: true
---
apiVersion: logging.banzaicloud.io/v1beta1
kind: ClusterFlow
metadata:
name: all-to-loki
namespace: cattle-logging-system
spec:
globalOutputRefs:
- loki
Harvester also routes Kubernetes audit logs and events through dedicated flows if configured.
← Backup, snapshot, restore and DR · Index · Harvester + Rancher integration →