openHAB

Helm chart for deploying openHAB on Kubernetes using the official openhab/openhab Docker image. openHAB is an open-source home automation platform that integrates with hundreds of smart home technologies — from Zigbee and Z-Wave to MQTT, KNX, and cloud services.

Single-instance only: openHAB does not support horizontal scaling. This chart enforces replicaCount: 1 and will fail fast if you attempt to set it higher, preventing accidental data corruption from concurrent writes to shared PVCs.

Key Features

StatefulSet workload for stable, predictable PVC attachment on every restart
Three persistent volumes (userdata, conf, addons) with independent sizing and storage class configuration
ConfigMap live-reload for sitemaps, things, and items — changes applied automatically within seconds, no pod restart required
Correct security context — fsGroup: 9001 only; runAsUser/runAsGroup intentionally unset so the entrypoint can bootstrap user creation before dropping privileges via gosu
Smart health probes via /rest/uuid (available in openHAB 4.x, no auth required) with a 10-minute startup window for OSGi bundle loading
Prometheus metrics via /rest/metrics/prometheus — pod annotations and ServiceMonitor supported
Optional Ingress with websocket annotation guidance for the /rest/events SSE endpoint
Optional Karaf SSH console for administrative access to the OSGi runtime
Fail-fast validation with clear error messages for common misconfigurations

Installation

HTTPS Repository

helm repo add helmforge https://repo.helmforge.dev
helm repo update
helm install my-openhab helmforge/openhab

OCI Registry

helm install my-openhab oci://ghcr.io/helmforgedev/helm/openhab --version 0.1.0

Quick Start

Minimal Deployment

Deploy openHAB with persistent storage and port-forward access:

helm install my-openhab helmforge/openhab
kubectl wait --namespace default pod \
  -l app.kubernetes.io/name=openhab \
  --for=condition=Ready --timeout=300s
kubectl port-forward svc/my-openhab 8080:8080

Open http://127.0.0.1:8080 and complete the first-boot admin setup wizard.

Custom Timezone and Resources

# values.yaml
env:
  TZ: 'Europe/Berlin'

resources:
  requests:
    cpu: 500m
    memory: 512Mi
  limits:
    cpu: 2000m
    memory: 2Gi

helm install my-openhab helmforge/openhab -f values.yaml

First Boot — Admin Setup

openHAB does not support injecting admin credentials via environment variables. On first launch, openHAB presents a setup wizard where you create the administrator account.

Steps:

Wait for the pod to be Ready (may take 60-120 seconds on first boot)
Navigate to the web UI (http://localhost:8080 if using port-forward)
Click Create an administrator account
Enter username and password
Complete the optional initial configuration steps

Credentials are stored persistently in /openhab/userdata/jsondb/auth.json and survive pod restarts as long as the userdata PVC exists.

Persistent Storage

openHAB requires three persistent directories. The chart creates a PVC for each by default.

Directory	PVC	Default Size	Contents
`/openhab/userdata`	`<release>-userdata`	`5Gi`	Runtime state, JSONDB, logs, persistence data
`/openhab/conf`	`<release>-conf`	`1Gi`	Items, things, rules, sitemaps, services
`/openhab/addons`	`<release>-addons`	`2Gi`	Drop-in JAR bindings/addons

Custom Sizes and Storage Class

persistence:
  userdata:
    size: 20Gi
    storageClass: 'fast-ssd'
  conf:
    size: 5Gi
    storageClass: 'fast-ssd'
  addons:
    size: 10Gi

Using Existing PVCs

persistence:
  userdata:
    existingClaim: my-openhab-userdata
  conf:
    existingClaim: my-openhab-conf
  addons:
    existingClaim: my-openhab-addons

ConfigMap Live Reload

This is the key differentiator of this chart. openHAB natively monitors /openhab/conf/ using a file watcher. Any change to .sitemap, .things, or .items files is applied automatically within 2-5 seconds — no pod restart required.

The chart mounts Kubernetes ConfigMaps into the conf PVC using subPath, so ConfigMap-managed files coexist with any existing files without overwriting them.

How It Works

Helm values  →  ConfigMap (K8s)  →  subPath mount  →  openHAB file watcher  →  Live reload (~2-5s)

Sitemaps

Sitemaps define the UI layout for openHAB’s BasicUI and HABPanel:

configMaps:
  sitemaps:
    enabled: true
    files:
      myhome.sitemap: |
        sitemap myhome label="My Home" {
          Frame label="Ground Floor" {
            Switch item=Light_GF_Corridor label="Corridor Light"
            Switch item=Light_GF_Kitchen  label="Kitchen Light"
            Text   item=Temperature_GF    label="Temperature [%.1f °C]"
          }
          Frame label="Climate" {
            Text item=Temperature_Outdoor label="Outdoor [%.1f °C]"
            Text item=Humidity_Outdoor   label="Humidity [%d %%]"
          }
        }

Things

Things define physical devices and their channels:

configMaps:
  things:
    enabled: true
    files:
      network.things: |
        Thing network:pingdevice:router [
          hostname="192.168.1.1",
          retry=1,
          timeout=5000,
          refreshInterval=60000
        ]
      mqtt.things: |
        Bridge mqtt:broker:mybroker [ host="mosquitto", port=1883, secure=false ] {
          Thing mqtt:topic:mysensor "Temperature Sensor" {
            Channels:
              Type number : temperature [ stateTopic="home/sensor/temperature" ]
              Type number : humidity    [ stateTopic="home/sensor/humidity" ]
          }
        }

Items

Items define logical entities visible in the UI and used in rules:

configMaps:
  items:
    enabled: true
    files:
      lights.items: |
        Switch Light_GF_Corridor "Corridor Light"  <light>
        Switch Light_GF_Kitchen  "Kitchen Light"   <light>
      climate.items: |
        Number:Temperature Temperature_GF       "Ground Floor [%.1f %unit%]" <temperature>
        Number:Temperature Temperature_Outdoor  "Outdoor [%.1f %unit%]"      <temperature>
        Number:Dimensionless Humidity_Outdoor   "Humidity [%d %%]"           <humidity>

Applying Changes

After updating ConfigMap values, run helm upgrade. openHAB picks up the changes automatically:

helm upgrade my-openhab helmforge/openhab -f values.yaml
# No restart needed — openHAB reloads configuration within seconds

Prometheus Metrics

openHAB exposes Prometheus-format metrics via the Metrics addon at:

GET /rest/metrics/prometheus   (port 8080, no authentication required)

Step 1 — Install the Metrics Addon

Install the addon once via the openHAB UI:

Settings → Add-on Store → Integrations → Metrics → Install

Or via the Karaf console (if enabled):

kubectl port-forward svc/my-openhab-karaf 8101:8101
ssh -p 8101 [email protected]
# Inside Karaf:
feature:install openhab-io-metrics

Step 2 — Enable Metrics in Chart Values

Mode 1: Pod annotations — works with any Prometheus that watches pod annotations:

metrics:
  enabled: true
  podAnnotations:
    enabled: true

This adds the following annotations to the pod:

prometheus.io/scrape: 'true'
prometheus.io/path: /rest/metrics/prometheus
prometheus.io/port: '8080'

Mode 2: ServiceMonitor — for Prometheus Operator / kube-prometheus-stack:

metrics:
  enabled: true
  podAnnotations:
    enabled: false
  serviceMonitor:
    enabled: true
    interval: 60s
    scrapeTimeout: 10s
    # Must match your Prometheus instance's serviceMonitorSelector labels
    additionalLabels:
      release: prometheus

Verify the Endpoint

kubectl port-forward svc/my-openhab 8080:8080
curl -s http://localhost:8080/rest/metrics/prometheus | head -20
# Expected: Prometheus text format with jvm_*, openhab_* metrics

Metrics Exposed

Category	Metrics
openHAB events	`openhab_events_total` per topic
Bundle states	`openhab_bundle_state` (32 = active)
Thing states	`openhab_thing_state` (online/offline)
Rule executions	`openhab_rule_runs_total`
Threadpool	`openhab_threadpool_*` (size, active, queue)
JVM memory	`jvm_memory_used_bytes`, `jvm_gc_pause_seconds`
JVM threads	`jvm_threads_*`, `jvm_classes_loaded`
Process	`process_cpu_usage`, `process_uptime_seconds`

Custom Relabelings (ServiceMonitor)

metrics:
  serviceMonitor:
    enabled: true
    relabelings:
      - sourceLabels: [__meta_kubernetes_pod_label_app_kubernetes_io_instance]
        targetLabel: instance
    metricRelabelings:
      - sourceLabels: [__name__]
        regex: 'jvm_.*'
        action: keep

Ingress

openHAB’s /rest/events endpoint uses Server-Sent Events (SSE), which requires long-lived HTTP connections. When using nginx Ingress, add the following annotations for proper support:

ingress:
  enabled: true
  ingressClassName: nginx
  annotations:
    nginx.ingress.kubernetes.io/proxy-read-timeout: '3600'
    nginx.ingress.kubernetes.io/proxy-send-timeout: '3600'
    nginx.ingress.kubernetes.io/proxy-http-version: '1.1'
    nginx.ingress.kubernetes.io/configuration-snippet: |
      proxy_set_header Upgrade $http_upgrade;
      proxy_set_header Connection "upgrade";
  hosts:
    - host: openhab.yourdomain.com
      paths:
        - path: /
          pathType: Prefix

With TLS (cert-manager)

ingress:
  enabled: true
  ingressClassName: nginx
  annotations:
    cert-manager.io/cluster-issuer: letsencrypt-prod
    nginx.ingress.kubernetes.io/proxy-read-timeout: '3600'
    nginx.ingress.kubernetes.io/proxy-send-timeout: '3600'
  hosts:
    - host: openhab.yourdomain.com
      paths:
        - path: /
          pathType: Prefix
  tls:
    - secretName: openhab-tls
      hosts:
        - openhab.yourdomain.com

Security

Security Context

openHAB runs as UID/GID 9001 by default (enforced by the official image). The chart configures this correctly out of the box:

podSecurityContext:
  runAsUser: 9001
  runAsGroup: 9001
  fsGroup: 9001 # Ensures PVC volumes are group-writable by 9001

securityContext:
  allowPrivilegeEscalation: false
  readOnlyRootFilesystem: false # openHAB writes to internal dirs at runtime
  capabilities:
    drop:
      - ALL

Why readOnlyRootFilesystem: false? openHAB (OSGi/Karaf) writes to several internal directories at runtime (/openhab/runtime/, /openhab/userdata/tmp/, /openhab/userdata/cache/). These cannot be relocated. Mount your persistent data on the three PVCs to ensure durability across restarts.

Karaf SSH Console

The Apache Karaf admin console is disabled by default. When enabled, always access it via kubectl port-forward — never expose port 8101 publicly:

karaf:
  enabled: true
  service:
    type: ClusterIP # Never NodePort or LoadBalancer
    port: 8101

kubectl port-forward svc/my-openhab-karaf 8101:8101
ssh -p 8101 [email protected]
# Default Karaf password: habopen

Admin Credentials Secret

For operational reference (documentation, tooling), you can store the admin credentials in a Kubernetes Secret:

admin:
  secretEnabled: true
  username: admin
  password: 'strongpassword' # Set via --set or external secret manager

Important: This Secret is for reference only. It does NOT automatically configure openHAB. You still need to complete the first-boot wizard with the same credentials.

Retrieve the stored password:

kubectl get secret my-openhab-admin \
  -o jsonpath="{.data.password}" | base64 --decode

Deployment Scenarios

Scenario 1: Minimal Home Lab

# Simple single-node home deployment
image:
  tag: '4.2.2'

env:
  TZ: 'Europe/Berlin'

persistence:
  userdata:
    size: 10Gi
  conf:
    size: 2Gi
  addons:
    size: 5Gi

Scenario 2: GitOps-Managed Configuration

Manage all openHAB configuration declaratively via Helm — ideal for teams or reproducible setups:

env:
  TZ: 'America/Sao_Paulo'

configMaps:
  sitemaps:
    enabled: true
    files:
      default.sitemap: |
        sitemap default label="openHAB" {
          Frame label="Overview" {
            Text item=gTemperature label="Temperature [%.1f °C]"
          }
        }
  things:
    enabled: true
    files:
      network.things: |
        Thing network:pingdevice:gateway [ hostname="192.168.1.1" ]
  items:
    enabled: true
    files:
      home.items: |
        Number:Temperature gTemperature "Temperature [%.1f %unit%]" <temperature>

persistence:
  userdata:
    size: 10Gi
  conf:
    size: 2Gi

Scenario 3: Full Production with Ingress and Metrics

ingress:
  enabled: true
  ingressClassName: nginx
  annotations:
    cert-manager.io/cluster-issuer: letsencrypt-prod
    nginx.ingress.kubernetes.io/proxy-read-timeout: '3600'
    nginx.ingress.kubernetes.io/proxy-send-timeout: '3600'
    nginx.ingress.kubernetes.io/proxy-http-version: '1.1'
    nginx.ingress.kubernetes.io/configuration-snippet: |
      proxy_set_header Upgrade $http_upgrade;
      proxy_set_header Connection "upgrade";
  hosts:
    - host: openhab.yourdomain.com
      paths:
        - path: /
          pathType: Prefix
  tls:
    - secretName: openhab-tls
      hosts:
        - openhab.yourdomain.com

admin:
  secretEnabled: true
  username: admin
  password: '' # Set via: --set admin.password=<value>

karaf:
  enabled: true

# Prometheus Operator (kube-prometheus-stack)
# Requires: openHAB Metrics addon installed via UI
metrics:
  enabled: true
  podAnnotations:
    enabled: false
  serviceMonitor:
    enabled: true
    interval: 60s
    additionalLabels:
      release: prometheus

env:
  TZ: 'Europe/Berlin'
  EXTRA_JAVA_OPTS: '-Xms512m -Xmx1536m'

persistence:
  userdata:
    size: 20Gi
    storageClass: 'fast-ssd'
  conf:
    size: 5Gi
    storageClass: 'fast-ssd'
  addons:
    size: 10Gi

resources:
  requests:
    cpu: 500m
    memory: 768Mi
  limits:
    cpu: 4000m
    memory: 3Gi

Automated Backup

The chart includes an optional CronJob that creates compressed archives of your openHAB data and uploads them to any S3-compatible object storage using the MinIO client (mc).

How It Works

The backup job runs as two containers sharing an emptyDir volume:

backup initContainer (alpine) — tars selected directories and writes the archive to /tmp
upload container (helmforge/mc) — picks up the archive and uploads it to S3

Both containers run as UID/GID 9001 to match openHAB’s PVC ownership.

Enabling Backup

backup:
  enabled: true
  schedule: '0 3 * * *' # Daily at 03:00 UTC

  s3:
    endpoint: 'https://minio.example.com'
    bucket: 'openhab-backups'
    prefix: 'prod'
    accessKey: 'AKIAEXAMPLE'
    secretKey: 'supersecretkey'

Using an Existing Secret

Avoid storing credentials in values by referencing a pre-created Secret:

kubectl create secret generic my-s3-creds \
  --from-literal=access-key=AKIAEXAMPLE \
  --from-literal=secret-key=supersecretkey \
  -n openhab

backup:
  enabled: true
  s3:
    endpoint: 'https://minio.example.com'
    bucket: 'openhab-backups'
    existingSecret: 'my-s3-creds'

What Gets Backed Up

Directory	Default	Description
`/openhab/userdata`	✅	JSONDB, persistence data, rules state
`/openhab/conf`	✅	Items, things, sitemaps, rules files

Always excluded from userdata: logs/, tmp/, cache/ (ephemeral — not needed for restore).

If you manage /openhab/conf via ConfigMaps (GitOps), you can skip it:

backup:
  include:
    userdata: true
    conf: false

S3 Compatibility

The uploader (helmforge/mc) is compatible with any S3-compatible service:

Provider	Endpoint format
MinIO	`https://minio.example.com`
AWS S3	`https://s3.amazonaws.com`
Cloudflare R2	`https://<account>.r2.cloudflarestorage.com`
Backblaze B2	`https://s3.<region>.backblazeb2.com`
DigitalOcean Spaces	`https://<region>.digitaloceanspaces.com`

Archive Naming

Archives follow this pattern: <archivePrefix>-backup-<YYYY-MM-DD-HHmmss>.tar.gz

Default: openhab-backup-2025-01-15-030000.tar.gz

Restore Process

Download the archive from your S3 bucket.
Scale down openHAB (required — it holds locks on its data):

kubectl scale statefulset my-openhab -n openhab --replicas=0

Extract via a temporary pod:

kubectl run restore --rm -it --image=alpine --restart=Never \
  --overrides='{"spec":{"volumes":[{"name":"ud","persistentVolumeClaim":{"claimName":"my-openhab-userdata"}}],"containers":[{"name":"r","image":"alpine","command":["sh"],"stdin":true,"tty":true,"volumeMounts":[{"name":"ud","mountPath":"/openhab/userdata"}]}]}}' \
  -- sh

# Inside the pod:
# tar -xzf /path/to/openhab-backup-<timestamp>.tar.gz -C /

Scale back to 1:

kubectl scale statefulset my-openhab -n openhab --replicas=1

Configuration Reference

Image

Parameter	Description	Default
`image.repository`	Image repository	`docker.io/openhab/openhab`
`image.tag`	Image tag	`4.2.2`
`image.pullPolicy`	Pull policy	`IfNotPresent`

Workload

Parameter	Description	Default
`replicaCount`	Must be `1` — clustering not supported	`1`
`podSecurityContext.runAsUser`	UID (required by openHAB image)	`9001`
`podSecurityContext.runAsGroup`	GID (required by openHAB image)	`9001`
`podSecurityContext.fsGroup`	fsGroup for PVC ownership	`9001`

Service

Parameter	Description	Default
`service.type`	Service type	`ClusterIP`
`service.port`	HTTP port	`8080`

Persistence

Parameter	Description	Default
`persistence.userdata.enabled`	Enable userdata PVC	`true`
`persistence.userdata.size`	PVC size	`5Gi`
`persistence.userdata.storageClass`	Storage class	`""`
`persistence.userdata.existingClaim`	Use existing PVC	`""`
`persistence.conf.enabled`	Enable conf PVC	`true`
`persistence.conf.size`	PVC size	`1Gi`
`persistence.addons.enabled`	Enable addons PVC	`true`
`persistence.addons.size`	PVC size	`2Gi`

ConfigMaps

Parameter	Description	Default
`configMaps.sitemaps.enabled`	Enable sitemaps ConfigMap	`false`
`configMaps.sitemaps.files`	Map of `filename → content`	`{}`
`configMaps.things.enabled`	Enable things ConfigMap	`false`
`configMaps.things.files`	Map of `filename → content`	`{}`
`configMaps.items.enabled`	Enable items ConfigMap	`false`
`configMaps.items.files`	Map of `filename → content`	`{}`

Metrics

Parameter	Description	Default
`metrics.enabled`	Enable Prometheus metrics	`false`
`metrics.podAnnotations.enabled`	Add `prometheus.io/*` pod annotations	`true`
`metrics.serviceMonitor.enabled`	Create ServiceMonitor	`false`
`metrics.serviceMonitor.namespace`	ServiceMonitor namespace	release namespace
`metrics.serviceMonitor.interval`	Scrape interval	`60s`
`metrics.serviceMonitor.scrapeTimeout`	Scrape timeout	`10s`
`metrics.serviceMonitor.additionalLabels`	Extra labels on ServiceMonitor	`{}`
`metrics.serviceMonitor.relabelings`	Relabeling rules	`[]`
`metrics.serviceMonitor.metricRelabelings`	Metric relabeling rules	`[]`

Environment

Parameter	Description	Default
`env.TZ`	Timezone	`UTC`
`env.EXTRA_JAVA_OPTS`	Extra JVM options	`""`
`env.OPENHAB_HTTP_PORT`	HTTP port	`8080`
`env.OPENHAB_HTTPS_PORT`	HTTPS port	`8443`

Optional Components

Parameter	Description	Default
`karaf.enabled`	Enable Karaf SSH console	`false`
`karaf.service.port`	Karaf SSH port	`8101`
`admin.secretEnabled`	Create admin credentials Secret	`false`
`admin.username`	Admin username (stored in Secret)	`admin`
`admin.password`	Admin password (stored in Secret)	`""`
`admin.existingSecret`	Use existing Secret	`""`

Backup

Parameter	Description	Default
`backup.enabled`	Enable automated backup CronJob	`false`
`backup.schedule`	Cron schedule	`0 3 * * *`
`backup.suspend`	Suspend the CronJob without deleting it	`false`
`backup.concurrencyPolicy`	CronJob concurrency policy	`Forbid`
`backup.successfulJobsHistoryLimit`	Successful job history to retain	`3`
`backup.failedJobsHistoryLimit`	Failed job history to retain	`3`
`backup.backoffLimit`	Job backoff limit	`1`
`backup.archivePrefix`	Archive filename prefix	`openhab`
`backup.include.userdata`	Back up `/openhab/userdata`	`true`
`backup.include.conf`	Back up `/openhab/conf`	`true`
`backup.images.utility.repository`	Backup utility image	`docker.io/library/alpine`
`backup.images.utility.tag`	Backup utility tag	`3.22`
`backup.images.uploader.repository`	S3 uploader image	`docker.io/helmforge/mc`
`backup.images.uploader.tag`	S3 uploader tag	`1.0.0`
`backup.resources`	Resource requests/limits for backup containers	`{}`
`backup.s3.endpoint`	S3-compatible endpoint URL	`""`
`backup.s3.bucket`	Target bucket name	`""`
`backup.s3.prefix`	Key prefix within the bucket	`openhab`
`backup.s3.accessKey`	S3 access key	`""`
`backup.s3.secretKey`	S3 secret key	`""`
`backup.s3.existingSecret`	Existing Secret name (keys: `access-key`, `secret-key`)	`""`

Troubleshooting

Pod stuck in Init state for several minutes

This is normal on first boot. openHAB loads all OSGi bundles at startup, which takes 60-120 seconds. The chart configures a startup probe with a 5-minute window. Wait patiently before investigating.

kubectl get pod -l app.kubernetes.io/name=openhab \
  -o jsonpath='{.items[0].status.containerStatuses[0].started}'

CrashLoopBackOff after PVC reuse

Incorrect permissions on a reused PVC will prevent openHAB from writing to its directories.

kubectl logs -l app.kubernetes.io/name=openhab --previous

# Fix permissions via a temporary pod
kubectl run fix-perms --image=busybox --restart=Never \
  --overrides='{"spec":{"volumes":[{"name":"data","persistentVolumeClaim":{"claimName":"my-openhab-userdata"}}],"containers":[{"name":"fix","image":"busybox","command":["chown","-R","9001:9001","/data"],"volumeMounts":[{"name":"data","mountPath":"/data"}]}]}}' \
  -- chown -R 9001:9001 /data

ConfigMap changes not reflected

Kubernetes syncs ConfigMaps to pods every ~60 seconds (kubelet --sync-frequency). After the file appears on disk, openHAB applies it within 2-5 seconds. If still not reflected after 2-3 minutes:

# Verify the file exists in the pod
kubectl exec -l app.kubernetes.io/name=openhab -- ls /openhab/conf/sitemaps/

# Check openHAB log for parsing errors
kubectl exec -l app.kubernetes.io/name=openhab -- \
  tail -50 /openhab/userdata/logs/openhab.log

403 Forbidden on web UI

The first-boot admin setup wizard has not been completed. Navigate to the UI root and create your administrator account.

Web UI works but real-time updates are broken

The /rest/events SSE endpoint requires long-lived connections. Verify the websocket annotations are present on your Ingress:

kubectl describe ingress my-openhab
# Look for: nginx.ingress.kubernetes.io/proxy-read-timeout: 3600

Prometheus metrics endpoint returns 404

The Metrics addon is not installed or not yet active. Install it via Settings → Add-on Store → Integrations → Metrics.

# Verify the endpoint manually
kubectl port-forward svc/my-openhab 8080:8080
curl -s http://localhost:8080/rest/metrics/prometheus | head -5

No metrics in Prometheus (annotations mode)

Verify the pod has the scrape annotations:

kubectl get pod -l app.kubernetes.io/name=openhab \
  -o jsonpath='{.items[0].metadata.annotations}' | jq .

ServiceMonitor not discovered by Prometheus

The additionalLabels on the ServiceMonitor must match the serviceMonitorSelector of your Prometheus instance:

kubectl get prometheus -o jsonpath='{.items[0].spec.serviceMonitorSelector}'
# Then set matching labels in metrics.serviceMonitor.additionalLabels

openHAB log shows bundle resolution errors on startup

Delete the OSGi bundle cache (rebuilt automatically on next start):

kubectl exec -l app.kubernetes.io/name=openhab -- \
  rm -rf /openhab/userdata/cache
kubectl rollout restart statefulset my-openhab

High memory usage

Tune the JVM heap. Keep resource limits above the -Xmx value to avoid OOM kills:

env:
  EXTRA_JAVA_OPTS: '-Xms256m -Xmx1024m'
resources:
  limits:
    memory: 1536Mi # must be > Xmx

Pod evicted due to storage pressure

The userdata volume accumulates logs over time. Increase the PVC size or configure log rotation:

persistence:
  userdata:
    size: 20Gi

Additional Resources

Mosquitto — MQTT broker for openHAB smart home integrations
MariaDB — Database backend for openHAB JDBC persistence

openHAB

Key Features

Installation

HTTPS Repository

OCI Registry

Quick Start

Minimal Deployment

Custom Timezone and Resources

First Boot — Admin Setup

Persistent Storage

Custom Sizes and Storage Class

Using Existing PVCs

ConfigMap Live Reload

How It Works

Sitemaps

Things

Items

Applying Changes

Prometheus Metrics

Step 1 — Install the Metrics Addon

Step 2 — Enable Metrics in Chart Values

Verify the Endpoint

Metrics Exposed

Custom Relabelings (ServiceMonitor)

Ingress

With TLS (cert-manager)

Security

Security Context

Karaf SSH Console

Admin Credentials Secret

Deployment Scenarios

Scenario 1: Minimal Home Lab

Scenario 2: GitOps-Managed Configuration

Scenario 3: Full Production with Ingress and Metrics

Automated Backup

How It Works

Enabling Backup

Using an Existing Secret

What Gets Backed Up

S3 Compatibility

Archive Naming

Restore Process

Configuration Reference

Image

Workload

Service

Persistence

ConfigMaps

Metrics

Environment

Optional Components

Backup

Troubleshooting

Pod stuck in Init state for several minutes

CrashLoopBackOff after PVC reuse

ConfigMap changes not reflected

403 Forbidden on web UI

Web UI works but real-time updates are broken

Prometheus metrics endpoint returns 404

No metrics in Prometheus (annotations mode)

ServiceMonitor not discovered by Prometheus

openHAB log shows bundle resolution errors on startup

High memory usage

Pod evicted due to storage pressure

Additional Resources

Related Charts