Version: Next

Configuration

This guide covers how to configure Bison for your specific environment and requirements.

Helm Chart Configuration

Bison is configured primarily through Helm values. You can customize the installation by providing a values.yaml file or using --set flags.

Key Configuration Parameters

Parameter	Description	Default	Example
`auth.enabled`	Enable authentication	`false`	`true`
`auth.admin.username`	Admin username	`admin`	`admin`
`auth.admin.password`	Admin password (must be non-default when `auth.enabled=true`)	`""`	`changeme`
`apiServer.replicaCount`	API server replicas (safe to raise: scheduler uses leader election)	`2`	`3`
`apiServer.image.repository`	API server image	`ghcr.io/supermarioyl/bison/api-server`	-
`apiServer.image.tag`	API server image tag	(chart appVersion)	`latest`
`apiServer.autoscaling.enabled`	Enable HPA for the API server (then `replicaCount` is ignored)	`false`	`true`
`apiServer.podDisruptionBudget.enabled`	Keep min replicas available during node drains	`false`	`true`
`webUI.replicaCount`	Web UI replicas	`2`	`3`
`webUI.image.repository`	Web UI image	`ghcr.io/supermarioyl/bison/web-ui`	-
`webUI.image.tag`	Web UI image tag	(chart appVersion)	`latest`
`webUI.autoscaling.enabled`	Enable HPA for the Web UI	`false`	`true`
`networkPolicy.enabled`	Restrict api-server ingress to web-ui pods + release namespace	`false`	`true`
`dependencies.opencost.apiUrl`	OpenCost API endpoint	`http://opencost.opencost.svc.cluster.local:9003`	Custom URL

Example Custom Values

Create a custom-values.yaml file:

# Authentication
auth:
  enabled: true
  admin:
    username: admin
    password: MySecurePassword123

# API Server
apiServer:
  replicaCount: 3
  image:
    tag: 0.0.12
  resources:
    requests:
      cpu: 200m
      memory: 256Mi
    limits:
      cpu: 1000m
      memory: 512Mi

# Web UI
webUI:
  replicaCount: 3
  image:
    tag: 0.0.12
  resources:
    requests:
      cpu: 100m
      memory: 128Mi
    limits:
      cpu: 500m
      memory: 256Mi

# OpenCost Integration
dependencies:
  opencost:
    enabled: true
    apiUrl: http://opencost.opencost.svc.cluster.local:9003

# Node Selection (optional)
nodeSelector:
  node-role.kubernetes.io/control-plane: ""

# Tolerations (optional)
tolerations:
  - key: node-role.kubernetes.io/control-plane
    operator: Exists
    effect: NoSchedule

Install with custom values:

helm install bison oci://ghcr.io/supermarioyl/charts/bison \
  --version 0.0.12 \
  --namespace bison-system \
  --create-namespace \
  --values custom-values.yaml

Billing Configuration

Billing settings are configured through the Web UI or API after installation.

Access Billing Configuration

Via Web UI:
- Navigate to Settings > Billing Configuration
- Set pricing for CPU, Memory, GPU, and other resources
- Configure currency and billing intervals

Via API:

curl -X POST http://localhost:8080/api/v1/billing/config \
  -H "Content-Type: application/json" \
  -d '{
    "enabled": true,
    "currency": "USD",
    "pricing": {
      "cpu": 0.05,
      "memory": 0.01,
      "nvidia.com/gpu": 2.50
    },
    "billingInterval": "hourly"
  }'

Billing Parameters

Parameter	Description	Example
`enabled`	Enable/disable billing	`true`
`currency`	Currency for billing	`USD`, `CNY`, `EUR`
`pricing.cpu`	CPU price per core-hour	`0.05`
`pricing.memory`	Memory price per GB-hour	`0.01`
`pricing["nvidia.com/gpu"]`	GPU price per GPU-hour	`2.50`
`billingInterval`	Billing aggregation period	`hourly`, `daily`
`lowBalanceThreshold`	Warning threshold (%)	`20`
`suspendThreshold`	Auto-suspend threshold (%)	`5`

Example Billing Configuration

{
  "enabled": true,
  "currency": "USD",
  "pricing": {
    "cpu": 0.05,
    "memory": 0.01,
    "nvidia.com/gpu": 2.50,
    "nvidia.com/mig-1g.5gb": 0.50,
    "nvidia.com/mig-2g.10gb": 1.00
  },
  "billingInterval": "hourly",
  "lowBalanceThreshold": 20,
  "suspendThreshold": 5,
  "alertChannels": ["webhook", "dingtalk"]
}

Team Configuration

Creating Teams

Teams can be created through the Web UI or API:

Via Web UI:

Navigate to Teams page
Click Create Team
Set team name, quota, and initial balance

Via API:

curl -X POST http://localhost:8080/api/v1/teams \
  -H "Content-Type: application/json" \
  -d '{
    "name": "ml-team",
    "description": "Machine Learning Team",
    "quota": {
      "cpu": "20",
      "memory": "64Gi",
      "nvidia.com/gpu": "4"
    },
    "balance": 1000.00
  }'

Team Quotas

Team quotas define resource limits:

quota:
  cpu: "20"              # 20 CPU cores
  memory: "64Gi"         # 64 GB RAM
  nvidia.com/gpu: "4"    # 4 GPUs
  storage: "500Gi"       # 500 GB storage

Team Balance Management

Set initial balance and configure auto-recharge:

{
  "balance": 1000.00,
  "autoRecharge": {
    "enabled": true,
    "amount": 500.00,
    "schedule": "monthly",
    "threshold": 100.00
  }
}

Alert Configuration

Configure multi-channel alerts for low balance and quota warnings.

Webhook Alerts

{
  "type": "webhook",
  "enabled": true,
  "url": "https://your-webhook-endpoint.com/alerts",
  "headers": {
    "Authorization": "Bearer YOUR_TOKEN"
  },
  "template": {
    "title": "Bison Alert",
    "message": "Team {{.TeamName}} balance is {{.Balance}}"
  }
}

DingTalk Alerts

{
  "type": "dingtalk",
  "enabled": true,
  "webhook": "https://oapi.dingtalk.com/robot/send?access_token=YOUR_TOKEN",
  "secret": "YOUR_SECRET"
}

WeChat Work Alerts

{
  "type": "wechat",
  "enabled": true,
  "corpid": "YOUR_CORP_ID",
  "corpsecret": "YOUR_CORP_SECRET",
  "agentid": 1000001
}

OpenCost Integration

Configure OpenCost connection:

Check OpenCost Connectivity

# Test OpenCost API
kubectl port-forward -n opencost svc/opencost 9003:9003
curl http://localhost:9003/healthz

# Test allocation API
curl http://localhost:9003/allocation/compute?window=1d

Update OpenCost URL

If OpenCost is deployed in a different namespace or with a different service name:

helm upgrade bison oci://ghcr.io/supermarioyl/charts/bison \
  --version 0.0.12 \
  --set dependencies.opencost.apiUrl=http://my-opencost.custom-namespace.svc:9003 \
  --set dependencies.opencost.enabled=true \
  --namespace bison-system

Authentication & OIDC

Enable authentication and integrate with your SSO provider:

Basic Authentication

auth:
  enabled: true
  admin:
    username: admin
    password: SecurePassword123

OIDC Integration

auth:
  enabled: true
  oidc:
    enabled: true
    issuerURL: https://your-oidc-provider.com
    clientID: bison-client-id
    clientSecret: your-client-secret
    redirectURL: https://bison.example.com/callback

Environment Variables

Additional configuration can be provided via environment variables:

Variable	Description	Default
`KUBECONFIG`	Path to kubeconfig file	In-cluster config
`OPENCOST_URL`	OpenCost API URL	`http://opencost.opencost.svc.cluster.local:9003`
`AUTH_ENABLED`	Enable authentication	`false`
`ADMIN_USERNAME`	Admin username	`admin`
`ADMIN_PASSWORD`	Admin password. Must not be `admin` when `AUTH_ENABLED=true` — the server refuses to start with the default.	`admin` (dev only)
`JWT_SECRET`	JWT signing key. Required (non-default) when `AUTH_ENABLED=true` — startup is refused otherwise.	(dev default)
`CORS_ALLOWED_ORIGINS`	Comma-separated CORS allowlist; empty allows all origins	(empty)
`LEADER_ELECTION_ENABLED`	Run the billing / auto-recharge / alert scheduler under a Kubernetes lease so it executes on exactly one replica (safe to scale `apiServer.replicaCount`)	`true`
`LOG_LEVEL`	Logging level	`info`

The billing interval is configured in the billing settings (hours), not via an environment variable.

Set environment variables in Helm values:

apiServer:
  env:
    - name: LOG_LEVEL
      value: debug
    - name: BILLING_INTERVAL
      value: 5m

Advanced Configuration

Custom Resource Pricing

Price any Kubernetes resource:

{
  "pricing": {
    "cpu": 0.05,
    "memory": 0.01,
    "nvidia.com/gpu": 2.50,
    "amd.com/gpu": 2.00,
    "ephemeral-storage": 0.001,
    "custom.io/fpga": 5.00
  }
}

Multi-Cluster Support

Deploy Bison in each cluster with shared billing:

# Cluster A
apiServer:
  clusterName: prod-us-west

# Cluster B
apiServer:
  clusterName: prod-us-east

Next Steps

User Guides - Learn how to use Bison
Architecture - Understand the system design
Features - Explore all capabilities

Helm Chart Configuration​

Key Configuration Parameters​

Example Custom Values​

Billing Configuration​

Access Billing Configuration​

Billing Parameters​

Example Billing Configuration​

Team Configuration​

Creating Teams​

Team Quotas​

Team Balance Management​

Alert Configuration​

Webhook Alerts​

DingTalk Alerts​

WeChat Work Alerts​

OpenCost Integration​

Check OpenCost Connectivity​

Update OpenCost URL​

Authentication & OIDC​

Basic Authentication​

OIDC Integration​

Environment Variables​

Advanced Configuration​

Custom Resource Pricing​

Multi-Cluster Support​

Next Steps​