Skip to main content
The Overlord chart deploys Bilibili’s Overlord, a high-performance cache proxy that supports both Memcached and Redis protocols. It provides features like connection pooling, request pipelining, and cluster management for cache systems.

​
What It Deploys

This chart deploys:
  • Overlord web interface for management
  • Overlord API server for proxy operations
  • Overlord scheduler for cluster coordination
  • Service for HTTP traffic
  • Optional Ingress or HTTPRoute for external access
  • ConfigMap or Secret for application configuration

​
Installation

helm repo add douban https://douban.github.io/charts/
helm install my-overlord douban/overlord -f values.yaml

​
Key Configuration Parameters

​
image.repository
string
default:"douz/overlord"
Overlord Docker image repository
​
image.tag
string
default:"latest"
Overlord image tag
​
web.replicaCount
number
default:"1"
Number of web interface replicas
​
config.apiserver
string
default:""
API server configuration (TOML format)
​
config.scheduler
string
default:""
Scheduler configuration (TOML format)
​
secretName
string
default:"overlord"
Name of the secret containing configuration
​
existingSecret
boolean
default:"false"
Use an existing secret instead of creating a new one
​
service.type
string
default:"ClusterIP"
Kubernetes service type
​
service.port
number
default:"80"
Service port for HTTP traffic
​
ingress.enabled
boolean
default:"false"
Enable Ingress resource for external access
​
httpRoute.enabled
boolean
default:"false"
Enable HTTPRoute resource for Gateway API routing

​
Example Configurations

​
Basic Deployment with Redis Backend

web:
  replicaCount: 2

apiserver:
  resources:
    limits:
      cpu: 1000m
      memory: 1Gi
    requests:
      cpu: 500m
      memory: 512Mi

scheduler:
  resources:
    limits:
      cpu: 500m
      memory: 512Mi
    requests:
      cpu: 250m
      memory: 256Mi

config:
  apiserver: |
    [cache]
    [cache.redis]
    [[cache.redis.cluster]]
    name = "redis-cluster"
    hash_method = "fnv1a_64"
    hash_distribution = "ketama"
    hash_tag = ""
    dial_timeout = 1000
    read_timeout = 1000
    write_timeout = 1000
    node_connections = 2
    ping_fail_limit = 3
    ping_auto_eject = true
    
    [[cache.redis.cluster.servers]]
    weight = 1
    addr = "redis-node-1:6379"
    
    [[cache.redis.cluster.servers]]
    weight = 1
    addr = "redis-node-2:6379"
    
    [[cache.redis.cluster.servers]]
    weight = 1
    addr = "redis-node-3:6379"

ingress:
  enabled: true
  className: "nginx"
  hosts:
    - host: overlord.example.com
      paths:
        - path: /
          pathType: Prefix

​
Memcached Cluster Configuration

config:
  apiserver: |
    [cache]
    [cache.memcache]
    [[cache.memcache.cluster]]
    name = "memcache-cluster"
    hash_method = "fnv1a_64"
    hash_distribution = "ketama"
    dial_timeout = 1000
    read_timeout = 1000
    write_timeout = 1000
    node_connections = 2
    
    [[cache.memcache.cluster.servers]]
    weight = 1
    addr = "memcached-1:11211"
    
    [[cache.memcache.cluster.servers]]
    weight = 1
    addr = "memcached-2:11211"
    
    [[cache.memcache.cluster.servers]]
    weight = 1
    addr = "memcached-3:11211"

​
Using Existing Secret for Configuration

existingSecret: true
secretName: my-overlord-config
Create the secret: 
kubectl create secret generic my-overlord-config \
  --from-file=apiserver.toml=./apiserver.toml \
  --from-file=scheduler.toml=./scheduler.toml

​
With Gateway API HTTPRoute

httpRoute:
  enabled: true
  gatewayRef:
    name: envoy-gateway-bundle
    namespace: envoy-gateway-system
  reuseIngressConfiguration: true

​
Scheduler Settings

config:
  scheduler: |
    [scheduler]
    interval = 60
    
    [scheduler.cluster]
    name = "main-cluster"
    
    [scheduler.storage]
    type = "etcd"
    endpoints = [
      "etcd-1:2379",
      "etcd-2:2379",
      "etcd-3:2379"
    ]
    dial_timeout = 5000
    request_timeout = 2000

​
Multiple Cache Clusters

config:
  apiserver: |
    [cache]
    
    # Redis Cluster
    [cache.redis]
    [[cache.redis.cluster]]
    name = "redis-main"
    hash_method = "fnv1a_64"
    hash_distribution = "ketama"
    dial_timeout = 1000
    read_timeout = 1000
    write_timeout = 1000
    node_connections = 4
    
    [[cache.redis.cluster.servers]]
    weight = 1
    addr = "redis-1:6379"
    
    [[cache.redis.cluster.servers]]
    weight = 1
    addr = "redis-2:6379"
    
    # Memcached Cluster
    [cache.memcache]
    [[cache.memcache.cluster]]
    name = "memcache-session"
    hash_method = "fnv1a_64"
    hash_distribution = "ketama"
    dial_timeout = 1000
    read_timeout = 1000
    write_timeout = 1000
    node_connections = 2
    
    [[cache.memcache.cluster.servers]]
    weight = 1
    addr = "memcached-1:11211"
    
    [[cache.memcache.cluster.servers]]
    weight = 1
    addr = "memcached-2:11211"

​
HTTPRoute with Custom Paths

httpRoute:
  enabled: true
  reuseIngressConfiguration: false
  gatewayRef:
    name: envoy-gateway-bundle
    namespace: envoy-gateway-system
  hostnames:
    - overlord.example.com
  extraPaths:
    - matches:
        - path:
            type: PathPrefix
            value: /api/v2
      backendRefs:
        - name: overlord-apiserver
          port: 8080

​
Resource Recommendations

​
Production Resource Limits

web:
  replicaCount: 2

apiserver:
  resources:
    limits:
      cpu: 2000m
      memory: 2Gi
    requests:
      cpu: 1000m
      memory: 1Gi

scheduler:
  resources:
    limits:
      cpu: 1000m
      memory: 1Gi
    requests:
      cpu: 500m
      memory: 512Mi

resources:
  limits:
    cpu: 500m
    memory: 512Mi
  requests:
    cpu: 250m
    memory: 256Mi

Build docs developers (and LLMs) love

Get started for freeTalk to us