// Deployment-Referenz
Helm chart: fleeger-agent
Helm chart for the Fleeger spoke-agent. Chart version 1.0.0-rc1 (appVersion 1.0.0-rc1).
What it deploys
A single spoke-agent that runs at a location, collects Proxmox / Kubernetes / iDRAC metrics read-only, and publishes them to the Fleeger hub over NATS. It is a single-replica Deployment (a second replica would double-report) with a small persistent volume for its bootstrapped credentials.
The agent talks to the hub over a single URL: it bootstraps over HTTPS, then does metrics and inventory over the encrypted NATS WebSocket the hub hands it (wss://<hub>/nats). So only the hub URL is required. See Agent modes: remote spoke vs. bundled core.
Install
You need the hub's URL and a one-time registration token from the UI (Locations → Onboard).
Put the token in a Secret first, in the namespace you install into. Paste it at the prompt: it is not echoed, and it
stays off every command line, where any local user could read it with ps
while the command runs:
printf 'Registration token: '; stty -echo; IFS= read -r FLEEGER_TOKEN; stty echo; echo
printf '%s' "$FLEEGER_TOKEN" | kubectl -n fleeger create secret generic fleeger-agent-token \
--from-file=registration-token=/dev/stdin
unset FLEEGER_TOKENThen install from a checkout, with the fleeger/spoke-agent:local image built from it and loaded into the nodes:
helm install fleeger-agent charts/fleeger-agent -n fleeger \
--set image.tag=local \
--set hub.url=https://fleeger.example.com \
--set existingSecret=fleeger-agent-tokenOr from the OCI registry (the project's GitLab container registry — replace registry.example.com/fleeger with your ${CI_REGISTRY_IMAGE}):
helm install fleeger-agent oci://registry.example.com/fleeger/charts/fleeger-agent \
--version 1.0.0-rc1 -n fleeger \
--set image.repository=registry.example.com/fleeger/spoke-agent \
--set hub.url=https://fleeger.example.com \
--set existingSecret=fleeger-agent-tokenLeave image.tag unset: a published chart runs the agent image its own release pipeline pushed (its appVersion), so each chart upgrade brings the matching agent with it. A tag you set is stored in the release and carried into every later upgrade.
--set registrationToken=<token> also works, but puts the token on helm's command line.
The registration token is single-use. Once the agent bootstraps, its NATS creds are written to the persistent volume and must survive restarts — losing them means re-onboarding. The token Secret is no longer read after that, so you can delete it. While they are there a new token is ignored: to onboard again (after offboarding, or a credential rotation), helm uninstall the release first, which deletes the volume; see re-onboarding.
Values reference
| Key | Default | Description |
|---|---|---|
image.repository |
fleeger/spoke-agent |
The whole image name without the tag. Set it for an OCI install, e.g. registry.example.com/fleeger/spoke-agent: the default names Docker Hub, where Fleeger publishes nothing. |
image.tag |
"" |
Empty means the chart's appVersion, the tag its release published; set local for an image built from a checkout. A tag set here stays pinned across upgrades until cleared with --set image.tag=. |
image.pullPolicy |
IfNotPresent |
|
imagePullSecrets |
[] |
Names of pull secrets for a private (e.g. GitLab) registry. |
hub.url |
"" |
Hub endpoint, e.g. https://fleeger.example.com. Required. |
hub.natsUrl |
"" |
Optional; only for static (non-bootstrap) mode. Normally learned from the hub at bootstrap. |
registrationToken |
"" |
One-time token from the UI. Stored in a Secret. |
existingSecret |
"" |
Instead of registrationToken, a Secret holding it under key registration-token. |
persistence.enabled |
true |
Persist bootstrapped NATS creds (+ .meta.json). |
persistence.size |
16Mi |
|
persistence.storageClass |
"" |
|
persistence.accessModes |
[ReadWriteOnce] |
|
persistence.mountPath |
/var/lib/fleeger |
Agent defaults FLEEGER_NATS_CREDS to <mountPath>/agent.creds. |
logLevel |
info |
|
pollIntervalSecs |
30 |
|
tmpSizeLimit |
2Gi |
Size of the writable /tmp emptyDir. The container's rootfs is read-only, so this is the agent's scratch space. |
resources.requests |
25m / 32Mi |
|
resources.limits.memory |
128Mi |
|
serviceAccount.create |
true |
|
serviceAccount.name |
"" |
|
serviceAccount.annotations |
{} |
|
rbac.create |
false |
Read-only ClusterRole (nodes/pods + metrics.k8s.io) for monitoring the very cluster the agent runs in via its ServiceAccount. Off by default — Proxmox / iDRAC monitoring needs no cluster RBAC. |
automountServiceAccountToken |
null |
Project a ServiceAccount token into the pod. Unset ⇒ follows rbac.create, so by default no token is mounted (the agent reaches the hub over NATS, not the API server). |
podAnnotations |
{} |
|
nodeSelector / tolerations / affinity |
{} / [] / {} |
Documentation
- Agent modes: remote spoke vs. bundled core — bootstrap vs static mode, core agent vs spoke agent, local collector.
- Deploying Fleeger on Kubernetes — the full deployment walkthrough.