Kubernetes

Prerequisites

Kubernetes 1.24+
kubectl configured for your cluster
Traefik running as your ingress controller (for IngressRoute support)
Traefik writing access logs to a known path on the node

Architecture

┌─────────────────────────────────────────────────┐
│ Kubernetes Cluster                              │
│                                                 │
│  ┌──────────────┐      ┌───────────────────┐   │
│  │    Agent      │◄─────│    Dashboard      │   │
│  │  (Go, :5000)  │      │  (Node.js, :3000) │   │
│  │               │      │                   │   │
│  │  HostPath:    │      │  PVC:             │   │
│  │  /var/log/    │      │  /data (SQLite)   │   │
│  │  traefik      │      │                   │   │
│  │  (read-only)  │      │                   │   │
│  └──────────────┘      └───────────────────┘   │
│         │                        │              │
│    ClusterIP               ClusterIP            │
│    svc:5000                svc:3000              │
│                                  │              │
│                          IngressRoute           │
│                     (traefik.example.com)        │
└─────────────────────────────────────────────────┘

The agent is internal-only (ClusterIP). The dashboard is exposed via a Traefik IngressRoute. The dashboard proxies all API calls to the agent.

Deployment

Clone and configure

git clone https://github.com/hhftechnology/traefik-log-dashboard.git
cd traefik-log-dashboard/k8s

Edit secret.yaml with your auth token:

# Generate a token
openssl rand -hex 32

Edit ingressroute.yaml with your domain and cert resolver.

If Traefik logs are not at /var/log/traefik, edit the hostPath in agent-deployment.yaml.

Apply all manifests

kubectl apply -k .

This creates:

traefik-dashboard namespace
Secret with auth token
ConfigMaps for agent and dashboard
PVCs for position tracking and SQLite
Agent Deployment + ClusterIP Service
Dashboard Deployment + ClusterIP Service
Traefik IngressRoute

Verify

kubectl -n traefik-dashboard get pods -w
kubectl -n traefik-dashboard logs -f deploy/traefik-agent
kubectl -n traefik-dashboard logs -f deploy/traefik-dashboard

Install with Helm

helm install traefik-log-dashboard charts/traefik-log-dashboard \
  --set auth.token=$(openssl rand -hex 32) \
  --set ingress.host=dashboard.yourdomain.com \
  --set agent.logHostPath=/var/log/traefik

Custom values

Create a my-values.yaml for your environment:

auth:
  token: "your-secret-token"

agent:
  logHostPath: /var/log/traefik
  resources:
    requests:
      cpu: 100m
      memory: 128Mi
    limits:
      cpu: 1000m
      memory: 1Gi

dashboard:
  config:
    agentName: "Production"
    maxLogsDisplay: "2000"

ingress:
  host: dashboard.example.com
  tls:
    certResolver: letsencrypt

helm install traefik-log-dashboard charts/traefik-log-dashboard -f my-values.yaml

Verify

kubectl -n traefik-dashboard get pods -w

Key Configuration

HostPath Volume

The agent reads Traefik logs via a hostPath volume mount. Set this to the directory where Traefik writes its access logs:

Method	Configuration
Plain manifests	Edit `hostPath.path` in `agent-deployment.yaml`
Helm	`--set agent.logHostPath=/your/traefik/log/path`

The agent pod must run on a node where Traefik writes logs. If Traefik runs as a DaemonSet, consider changing the agent from a Deployment to a DaemonSet as well.

Authentication

Both the agent and dashboard use the same shared token for API authentication. The token is stored in a Kubernetes Secret (traefik-dashboard-auth) and injected via environment variables.

# Generate a secure token
openssl rand -hex 32

Persistence

Component	Mount Path	Purpose	Default Size
Agent	`/data`	Position file tracking	64Mi
Dashboard	`/data`	SQLite database	1Gi

Multi-Agent

To monitor multiple Traefik instances, add additional agent environment variables in the dashboard ConfigMap:

data:
  AGENT_1_NAME: "Production"
  AGENT_1_URL: "http://traefik-agent:5000"
  AGENT_2_NAME: "Staging"
  AGENT_2_URL: "http://staging-agent.other-namespace:5000"

Port Forwarding (Quick Access)

For testing without an IngressRoute:

kubectl -n traefik-dashboard port-forward svc/traefik-dashboard 3000:3000
# Open http://localhost:3000

Troubleshooting

Symptom	Check
Agent pod `CrashLoopBackOff`	Verify `hostPath` exists on the node and contains log files
Dashboard init container stuck	Agent not healthy — check agent logs
No logs showing	Verify Traefik is writing JSON logs to the configured path
IngressRoute not working	Ensure Traefik CRDs are installed (`kubectl get crd ingressroutes.traefik.io`)

Next Steps

Docker Compose Deployment — Simpler alternative for non-K8s environments
Configuration — Detailed configuration options
Environment Variables — All available env vars

On this page