wangbo
7786692d32
将 Worker 发现、路由画像、容量与执行传输抽象为平台无关接口，新增 Kubernetes 和静态容量适配器，并以 shadow 模式接入生产配置。 实现网络与容量评分、路由防抖、池队列、同步 Worker 租约、一次性执行令牌，以及提交状态不明时禁止重复分配的安全语义。 新增 0105 兼容迁移、管理接口、指标、OpenAPI 和回归测试。已执行全量 Go 测试、go vet、OpenAPI、迁移安全、Compose 与 Kustomize 验证。
1154 lines
33 KiB
YAML
1154 lines
33 KiB
YAML
apiVersion: apps/v1
|
|
kind: Deployment
|
|
metadata:
|
|
name: easyai-api-ningbo
|
|
labels:
|
|
app.kubernetes.io/name: easyai-api
|
|
app.kubernetes.io/component: api
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
spec:
|
|
replicas: 1
|
|
strategy:
|
|
type: RollingUpdate
|
|
rollingUpdate:
|
|
maxSurge: 1
|
|
maxUnavailable: 0
|
|
selector:
|
|
matchLabels:
|
|
app.kubernetes.io/name: easyai-api
|
|
easyai.io/site: ningbo
|
|
template:
|
|
metadata:
|
|
labels:
|
|
app.kubernetes.io/name: easyai-api
|
|
app.kubernetes.io/component: api
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
easyai.io/site: ningbo
|
|
spec:
|
|
serviceAccountName: easyai-ai-gateway
|
|
terminationGracePeriodSeconds: 90
|
|
nodeSelector:
|
|
easyai.io/site: ningbo
|
|
easyai.io/workload: "true"
|
|
securityContext:
|
|
runAsNonRoot: true
|
|
runAsUser: 10001
|
|
runAsGroup: 10001
|
|
fsGroup: 10001
|
|
fsGroupChangePolicy: OnRootMismatch
|
|
seccompProfile:
|
|
type: RuntimeDefault
|
|
containers:
|
|
- name: api
|
|
image: easyai-api
|
|
imagePullPolicy: IfNotPresent
|
|
envFrom:
|
|
- configMapRef:
|
|
name: easyai-ai-gateway-config
|
|
- secretRef:
|
|
name: easyai-ai-gateway-runtime
|
|
env:
|
|
- name: GOMEMLIMIT
|
|
value: 1536MiB
|
|
- name: AI_GATEWAY_PROCESS_ROLE
|
|
value: api
|
|
- name: AI_GATEWAY_ASYNC_QUEUE_WORKER_ENABLED
|
|
value: "false"
|
|
- name: AI_GATEWAY_ASYNC_ADMISSION_DISPATCHER_ENABLED
|
|
value: "true"
|
|
- name: AI_GATEWAY_DATABASE_MAX_CONNS
|
|
value: "32"
|
|
- name: AI_GATEWAY_DATABASE_CRITICAL_MAX_CONNS
|
|
value: "4"
|
|
- name: AI_GATEWAY_DATABASE_RIVER_MAX_CONNS
|
|
value: "4"
|
|
- name: AI_GATEWAY_DATABASE_MIN_IDLE_CONNS
|
|
value: "4"
|
|
- name: AI_GATEWAY_DATABASE_MAX_CONN_IDLE_SECONDS
|
|
value: "30"
|
|
- name: AI_GATEWAY_MEDIA_REQUEST_CONCURRENCY
|
|
value: "16"
|
|
- name: AI_GATEWAY_MEDIA_MATERIALIZATION_CONCURRENCY
|
|
value: "8"
|
|
- name: AI_GATEWAY_MEDIA_IMAGE_NORMALIZATION_CONCURRENCY
|
|
value: "2"
|
|
- name: AI_GATEWAY_MEDIA_OSS_ACCESS_KEY_ID
|
|
valueFrom:
|
|
secretKeyRef:
|
|
name: easyai-oss-backup
|
|
key: ACCESS_KEY_ID
|
|
- name: AI_GATEWAY_MEDIA_OSS_ACCESS_KEY_SECRET
|
|
valueFrom:
|
|
secretKeyRef:
|
|
name: easyai-oss-backup
|
|
key: ACCESS_SECRET_KEY
|
|
- name: POD_NAMESPACE
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: metadata.namespace
|
|
- name: POD_NAME
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: metadata.name
|
|
- name: POD_UID
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: metadata.uid
|
|
- name: POD_IP
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: status.podIP
|
|
- name: EASYAI_SITE
|
|
value: ningbo
|
|
ports:
|
|
- name: http
|
|
containerPort: 8088
|
|
startupProbe:
|
|
httpGet:
|
|
path: /api/v1/healthz
|
|
port: http
|
|
periodSeconds: 5
|
|
timeoutSeconds: 3
|
|
failureThreshold: 120
|
|
readinessProbe:
|
|
httpGet:
|
|
path: /api/v1/readyz
|
|
port: http
|
|
initialDelaySeconds: 3
|
|
periodSeconds: 5
|
|
timeoutSeconds: 3
|
|
failureThreshold: 6
|
|
livenessProbe:
|
|
httpGet:
|
|
path: /api/v1/healthz
|
|
port: http
|
|
initialDelaySeconds: 15
|
|
periodSeconds: 10
|
|
timeoutSeconds: 3
|
|
failureThreshold: 6
|
|
lifecycle:
|
|
preStop:
|
|
exec:
|
|
command: ["/bin/sh", "-c", "sleep 10"]
|
|
resources:
|
|
requests:
|
|
cpu: 250m
|
|
memory: 512Mi
|
|
limits:
|
|
cpu: "2"
|
|
memory: 2Gi
|
|
securityContext:
|
|
allowPrivilegeEscalation: false
|
|
capabilities:
|
|
drop: ["ALL"]
|
|
readOnlyRootFilesystem: true
|
|
volumeMounts:
|
|
- name: application-data
|
|
mountPath: /app/data
|
|
- name: temporary-files
|
|
mountPath: /tmp
|
|
volumes:
|
|
- name: application-data
|
|
emptyDir:
|
|
sizeLimit: 5Gi
|
|
- name: temporary-files
|
|
emptyDir:
|
|
sizeLimit: 1Gi
|
|
---
|
|
apiVersion: apps/v1
|
|
kind: Deployment
|
|
metadata:
|
|
name: easyai-capacity-controller
|
|
labels:
|
|
app.kubernetes.io/name: easyai-capacity-controller
|
|
app.kubernetes.io/component: capacity-controller
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
spec:
|
|
replicas: 1
|
|
strategy:
|
|
type: RollingUpdate
|
|
rollingUpdate:
|
|
maxSurge: 1
|
|
maxUnavailable: 1
|
|
selector:
|
|
matchLabels:
|
|
app.kubernetes.io/name: easyai-capacity-controller
|
|
template:
|
|
metadata:
|
|
labels:
|
|
app.kubernetes.io/name: easyai-capacity-controller
|
|
app.kubernetes.io/component: capacity-controller
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
spec:
|
|
serviceAccountName: easyai-capacity-controller
|
|
terminationGracePeriodSeconds: 30
|
|
nodeSelector:
|
|
easyai.io/site: ningbo
|
|
easyai.io/workload: "true"
|
|
securityContext:
|
|
runAsNonRoot: true
|
|
runAsUser: 10001
|
|
runAsGroup: 10001
|
|
fsGroup: 10001
|
|
seccompProfile:
|
|
type: RuntimeDefault
|
|
containers:
|
|
- name: capacity-controller
|
|
image: easyai-api
|
|
imagePullPolicy: IfNotPresent
|
|
envFrom:
|
|
- configMapRef:
|
|
name: easyai-ai-gateway-config
|
|
env:
|
|
- name: AI_GATEWAY_PROCESS_ROLE
|
|
value: capacity-controller
|
|
- name: AI_GATEWAY_DATABASE_URL
|
|
valueFrom:
|
|
secretKeyRef:
|
|
name: easyai-ai-gateway-runtime
|
|
key: AI_GATEWAY_DATABASE_URL
|
|
- name: AI_GATEWAY_DATABASE_MAX_CONNS
|
|
value: "4"
|
|
- name: AI_GATEWAY_DATABASE_CRITICAL_MAX_CONNS
|
|
value: "0"
|
|
- name: AI_GATEWAY_DATABASE_RIVER_MAX_CONNS
|
|
value: "0"
|
|
- name: AI_GATEWAY_DATABASE_MIN_IDLE_CONNS
|
|
value: "1"
|
|
- name: AI_GATEWAY_WORKER_DATABASE_MAX_CONNS
|
|
value: "32"
|
|
- name: AI_GATEWAY_POSTGRES_NON_WORKER_CONNECTION_BUDGET
|
|
value: "36"
|
|
- name: POD_NAMESPACE
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: metadata.namespace
|
|
- name: POD_NAME
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: metadata.name
|
|
ports:
|
|
- name: health
|
|
containerPort: 8088
|
|
startupProbe:
|
|
httpGet:
|
|
path: /healthz
|
|
port: health
|
|
periodSeconds: 5
|
|
timeoutSeconds: 3
|
|
failureThreshold: 24
|
|
readinessProbe:
|
|
httpGet:
|
|
path: /readyz
|
|
port: health
|
|
periodSeconds: 10
|
|
timeoutSeconds: 3
|
|
failureThreshold: 6
|
|
livenessProbe:
|
|
httpGet:
|
|
path: /healthz
|
|
port: health
|
|
periodSeconds: 10
|
|
timeoutSeconds: 3
|
|
failureThreshold: 6
|
|
resources:
|
|
requests:
|
|
cpu: 50m
|
|
memory: 64Mi
|
|
limits:
|
|
cpu: 200m
|
|
memory: 256Mi
|
|
securityContext:
|
|
allowPrivilegeEscalation: false
|
|
capabilities:
|
|
drop: ["ALL"]
|
|
readOnlyRootFilesystem: true
|
|
volumeMounts:
|
|
- name: temporary-files
|
|
mountPath: /tmp
|
|
volumes:
|
|
- name: temporary-files
|
|
emptyDir:
|
|
sizeLimit: 64Mi
|
|
---
|
|
apiVersion: apps/v1
|
|
kind: Deployment
|
|
metadata:
|
|
name: easyai-api-hongkong
|
|
labels:
|
|
app.kubernetes.io/name: easyai-api
|
|
app.kubernetes.io/component: api
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
spec:
|
|
replicas: 1
|
|
strategy:
|
|
type: RollingUpdate
|
|
rollingUpdate:
|
|
maxSurge: 1
|
|
maxUnavailable: 0
|
|
selector:
|
|
matchLabels:
|
|
app.kubernetes.io/name: easyai-api
|
|
easyai.io/site: hongkong
|
|
template:
|
|
metadata:
|
|
labels:
|
|
app.kubernetes.io/name: easyai-api
|
|
app.kubernetes.io/component: api
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
easyai.io/site: hongkong
|
|
spec:
|
|
serviceAccountName: easyai-ai-gateway
|
|
terminationGracePeriodSeconds: 90
|
|
nodeSelector:
|
|
easyai.io/site: hongkong
|
|
easyai.io/workload: "true"
|
|
securityContext:
|
|
runAsNonRoot: true
|
|
runAsUser: 10001
|
|
runAsGroup: 10001
|
|
fsGroup: 10001
|
|
fsGroupChangePolicy: OnRootMismatch
|
|
seccompProfile:
|
|
type: RuntimeDefault
|
|
containers:
|
|
- name: api
|
|
image: easyai-api
|
|
imagePullPolicy: IfNotPresent
|
|
envFrom:
|
|
- configMapRef:
|
|
name: easyai-ai-gateway-config
|
|
- secretRef:
|
|
name: easyai-ai-gateway-runtime
|
|
env:
|
|
- name: GOMEMLIMIT
|
|
value: 1536MiB
|
|
- name: AI_GATEWAY_PROCESS_ROLE
|
|
value: api
|
|
- name: AI_GATEWAY_ASYNC_QUEUE_WORKER_ENABLED
|
|
value: "false"
|
|
- name: AI_GATEWAY_ASYNC_ADMISSION_DISPATCHER_ENABLED
|
|
value: "true"
|
|
- name: AI_GATEWAY_DATABASE_MAX_CONNS
|
|
value: "32"
|
|
- name: AI_GATEWAY_DATABASE_CRITICAL_MAX_CONNS
|
|
value: "4"
|
|
- name: AI_GATEWAY_DATABASE_RIVER_MAX_CONNS
|
|
value: "4"
|
|
- name: AI_GATEWAY_DATABASE_MIN_IDLE_CONNS
|
|
value: "4"
|
|
- name: AI_GATEWAY_DATABASE_MAX_CONN_IDLE_SECONDS
|
|
value: "30"
|
|
- name: AI_GATEWAY_MEDIA_REQUEST_CONCURRENCY
|
|
value: "16"
|
|
- name: AI_GATEWAY_MEDIA_MATERIALIZATION_CONCURRENCY
|
|
value: "8"
|
|
- name: AI_GATEWAY_MEDIA_IMAGE_NORMALIZATION_CONCURRENCY
|
|
value: "2"
|
|
- name: AI_GATEWAY_MEDIA_OSS_ACCESS_KEY_ID
|
|
valueFrom:
|
|
secretKeyRef:
|
|
name: easyai-oss-backup
|
|
key: ACCESS_KEY_ID
|
|
- name: AI_GATEWAY_MEDIA_OSS_ACCESS_KEY_SECRET
|
|
valueFrom:
|
|
secretKeyRef:
|
|
name: easyai-oss-backup
|
|
key: ACCESS_SECRET_KEY
|
|
- name: POD_NAMESPACE
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: metadata.namespace
|
|
- name: POD_NAME
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: metadata.name
|
|
- name: POD_UID
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: metadata.uid
|
|
- name: POD_IP
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: status.podIP
|
|
- name: EASYAI_SITE
|
|
value: hongkong
|
|
ports:
|
|
- name: http
|
|
containerPort: 8088
|
|
startupProbe:
|
|
httpGet:
|
|
path: /api/v1/healthz
|
|
port: http
|
|
periodSeconds: 5
|
|
timeoutSeconds: 3
|
|
failureThreshold: 120
|
|
readinessProbe:
|
|
httpGet:
|
|
path: /api/v1/readyz
|
|
port: http
|
|
initialDelaySeconds: 3
|
|
periodSeconds: 5
|
|
timeoutSeconds: 3
|
|
failureThreshold: 6
|
|
livenessProbe:
|
|
httpGet:
|
|
path: /api/v1/healthz
|
|
port: http
|
|
initialDelaySeconds: 15
|
|
periodSeconds: 10
|
|
timeoutSeconds: 3
|
|
failureThreshold: 6
|
|
lifecycle:
|
|
preStop:
|
|
exec:
|
|
command: ["/bin/sh", "-c", "sleep 10"]
|
|
resources:
|
|
requests:
|
|
cpu: 250m
|
|
memory: 512Mi
|
|
limits:
|
|
cpu: "2"
|
|
memory: 2Gi
|
|
securityContext:
|
|
allowPrivilegeEscalation: false
|
|
capabilities:
|
|
drop: ["ALL"]
|
|
readOnlyRootFilesystem: true
|
|
volumeMounts:
|
|
- name: application-data
|
|
mountPath: /app/data
|
|
- name: temporary-files
|
|
mountPath: /tmp
|
|
volumes:
|
|
- name: application-data
|
|
emptyDir:
|
|
sizeLimit: 5Gi
|
|
- name: temporary-files
|
|
emptyDir:
|
|
sizeLimit: 1Gi
|
|
---
|
|
apiVersion: apps/v1
|
|
kind: Deployment
|
|
metadata:
|
|
name: easyai-worker-ningbo
|
|
labels:
|
|
app.kubernetes.io/name: easyai-worker
|
|
app.kubernetes.io/component: worker
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
spec:
|
|
replicas: 1
|
|
strategy:
|
|
type: RollingUpdate
|
|
rollingUpdate:
|
|
maxSurge: 0
|
|
maxUnavailable: 1
|
|
selector:
|
|
matchLabels:
|
|
app.kubernetes.io/name: easyai-worker
|
|
easyai.io/site: ningbo
|
|
template:
|
|
metadata:
|
|
labels:
|
|
app.kubernetes.io/name: easyai-worker
|
|
app.kubernetes.io/component: worker
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
easyai.io/site: ningbo
|
|
spec:
|
|
serviceAccountName: easyai-ai-gateway
|
|
terminationGracePeriodSeconds: 90
|
|
affinity:
|
|
podAntiAffinity:
|
|
requiredDuringSchedulingIgnoredDuringExecution:
|
|
- topologyKey: kubernetes.io/hostname
|
|
labelSelector:
|
|
matchLabels:
|
|
app.kubernetes.io/name: easyai-worker
|
|
easyai.io/site: ningbo
|
|
topologySpreadConstraints:
|
|
- maxSkew: 1
|
|
topologyKey: kubernetes.io/hostname
|
|
whenUnsatisfiable: DoNotSchedule
|
|
labelSelector:
|
|
matchLabels:
|
|
app.kubernetes.io/name: easyai-worker
|
|
easyai.io/site: ningbo
|
|
nodeSelector:
|
|
easyai.io/site: ningbo
|
|
easyai.io/workload: "true"
|
|
easyai.io/worker: "true"
|
|
tolerations:
|
|
- key: easyai.io/worker-only
|
|
operator: Equal
|
|
value: "true"
|
|
effect: NoSchedule
|
|
securityContext:
|
|
runAsNonRoot: true
|
|
runAsUser: 10001
|
|
runAsGroup: 10001
|
|
fsGroup: 10001
|
|
fsGroupChangePolicy: OnRootMismatch
|
|
seccompProfile:
|
|
type: RuntimeDefault
|
|
containers:
|
|
- name: worker
|
|
image: easyai-api
|
|
imagePullPolicy: IfNotPresent
|
|
envFrom:
|
|
- configMapRef:
|
|
name: easyai-ai-gateway-config
|
|
- secretRef:
|
|
name: easyai-ai-gateway-runtime
|
|
env:
|
|
- name: AI_GATEWAY_PROCESS_ROLE
|
|
value: worker
|
|
- name: AI_GATEWAY_ASYNC_QUEUE_WORKER_ENABLED
|
|
value: "true"
|
|
- name: AI_GATEWAY_ASYNC_ADMISSION_DISPATCHER_ENABLED
|
|
value: "false"
|
|
- name: AI_GATEWAY_DATABASE_MAX_CONNS
|
|
value: "32"
|
|
- name: AI_GATEWAY_DATABASE_CRITICAL_MAX_CONNS
|
|
value: "4"
|
|
- name: AI_GATEWAY_DATABASE_RIVER_MAX_CONNS
|
|
value: "8"
|
|
- name: AI_GATEWAY_DATABASE_MIN_IDLE_CONNS
|
|
value: "4"
|
|
- name: AI_GATEWAY_DATABASE_MAX_CONN_IDLE_SECONDS
|
|
value: "30"
|
|
- name: AI_GATEWAY_ASYNC_WORKER_INSTANCE_HARD_LIMIT
|
|
value: "24"
|
|
- name: AI_GATEWAY_ASYNC_ADMISSION_MICROBATCH_SIZE
|
|
value: "8"
|
|
- name: AI_GATEWAY_MEDIA_MATERIALIZATION_CONCURRENCY
|
|
value: "24"
|
|
- name: AI_GATEWAY_MEDIA_REQUEST_CONCURRENCY
|
|
value: "24"
|
|
- name: AI_GATEWAY_MEDIA_IMAGE_NORMALIZATION_CONCURRENCY
|
|
value: "2"
|
|
- name: AI_GATEWAY_MEDIA_OSS_ACCESS_KEY_ID
|
|
valueFrom:
|
|
secretKeyRef:
|
|
name: easyai-oss-backup
|
|
key: ACCESS_KEY_ID
|
|
- name: AI_GATEWAY_MEDIA_OSS_ACCESS_KEY_SECRET
|
|
valueFrom:
|
|
secretKeyRef:
|
|
name: easyai-oss-backup
|
|
key: ACCESS_SECRET_KEY
|
|
- name: POD_NAMESPACE
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: metadata.namespace
|
|
- name: POD_NAME
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: metadata.name
|
|
- name: POD_UID
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: metadata.uid
|
|
- name: POD_IP
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: status.podIP
|
|
- name: AI_GATEWAY_WORKER_ID
|
|
value: pool-cn-east-1-worker
|
|
- name: AI_GATEWAY_WORKER_ADVERTISE_ENDPOINT
|
|
value: http://$(POD_IP):8088
|
|
- name: AI_GATEWAY_ORCHESTRATOR_INSTANCE_REF
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: metadata.name
|
|
- name: AI_GATEWAY_EXECUTION_POOL_ID
|
|
value: pool-cn-east-1
|
|
- name: AI_GATEWAY_EXECUTION_POOL_LABELS
|
|
value: '{"region":"cn-east","zone":"ningbo","provider":"self-hosted","egress_profile":"direct"}'
|
|
ports:
|
|
- name: health
|
|
containerPort: 8088
|
|
startupProbe:
|
|
httpGet:
|
|
path: /healthz
|
|
port: health
|
|
periodSeconds: 5
|
|
timeoutSeconds: 3
|
|
failureThreshold: 120
|
|
readinessProbe:
|
|
httpGet:
|
|
path: /readyz
|
|
port: health
|
|
initialDelaySeconds: 3
|
|
periodSeconds: 5
|
|
timeoutSeconds: 3
|
|
failureThreshold: 6
|
|
livenessProbe:
|
|
httpGet:
|
|
path: /healthz
|
|
port: health
|
|
initialDelaySeconds: 15
|
|
periodSeconds: 10
|
|
timeoutSeconds: 3
|
|
failureThreshold: 6
|
|
lifecycle:
|
|
preStop:
|
|
exec:
|
|
command: ["/bin/sh", "-c", "sleep 10"]
|
|
resources:
|
|
requests:
|
|
cpu: 500m
|
|
memory: 1536Mi
|
|
limits:
|
|
cpu: "2"
|
|
memory: 2Gi
|
|
securityContext:
|
|
allowPrivilegeEscalation: false
|
|
capabilities:
|
|
drop: ["ALL"]
|
|
readOnlyRootFilesystem: true
|
|
volumeMounts:
|
|
- name: application-data
|
|
mountPath: /app/data
|
|
- name: temporary-files
|
|
mountPath: /tmp
|
|
volumes:
|
|
- name: application-data
|
|
emptyDir:
|
|
sizeLimit: 5Gi
|
|
- name: temporary-files
|
|
emptyDir:
|
|
sizeLimit: 1Gi
|
|
---
|
|
apiVersion: apps/v1
|
|
kind: Deployment
|
|
metadata:
|
|
name: easyai-worker-hongkong
|
|
labels:
|
|
app.kubernetes.io/name: easyai-worker
|
|
app.kubernetes.io/component: worker
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
spec:
|
|
replicas: 1
|
|
strategy:
|
|
type: RollingUpdate
|
|
rollingUpdate:
|
|
maxSurge: 0
|
|
maxUnavailable: 1
|
|
selector:
|
|
matchLabels:
|
|
app.kubernetes.io/name: easyai-worker
|
|
easyai.io/site: hongkong
|
|
template:
|
|
metadata:
|
|
labels:
|
|
app.kubernetes.io/name: easyai-worker
|
|
app.kubernetes.io/component: worker
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
easyai.io/site: hongkong
|
|
spec:
|
|
serviceAccountName: easyai-ai-gateway
|
|
terminationGracePeriodSeconds: 90
|
|
affinity:
|
|
podAntiAffinity:
|
|
requiredDuringSchedulingIgnoredDuringExecution:
|
|
- topologyKey: kubernetes.io/hostname
|
|
labelSelector:
|
|
matchLabels:
|
|
app.kubernetes.io/name: easyai-worker
|
|
easyai.io/site: hongkong
|
|
topologySpreadConstraints:
|
|
- maxSkew: 1
|
|
topologyKey: kubernetes.io/hostname
|
|
whenUnsatisfiable: DoNotSchedule
|
|
labelSelector:
|
|
matchLabels:
|
|
app.kubernetes.io/name: easyai-worker
|
|
easyai.io/site: hongkong
|
|
nodeSelector:
|
|
easyai.io/site: hongkong
|
|
easyai.io/workload: "true"
|
|
easyai.io/worker: "true"
|
|
tolerations:
|
|
- key: easyai.io/worker-only
|
|
operator: Equal
|
|
value: "true"
|
|
effect: NoSchedule
|
|
securityContext:
|
|
runAsNonRoot: true
|
|
runAsUser: 10001
|
|
runAsGroup: 10001
|
|
fsGroup: 10001
|
|
fsGroupChangePolicy: OnRootMismatch
|
|
seccompProfile:
|
|
type: RuntimeDefault
|
|
containers:
|
|
- name: worker
|
|
image: easyai-api
|
|
imagePullPolicy: IfNotPresent
|
|
envFrom:
|
|
- configMapRef:
|
|
name: easyai-ai-gateway-config
|
|
- secretRef:
|
|
name: easyai-ai-gateway-runtime
|
|
env:
|
|
- name: AI_GATEWAY_PROCESS_ROLE
|
|
value: worker
|
|
- name: AI_GATEWAY_ASYNC_QUEUE_WORKER_ENABLED
|
|
value: "true"
|
|
- name: AI_GATEWAY_ASYNC_ADMISSION_DISPATCHER_ENABLED
|
|
value: "false"
|
|
- name: AI_GATEWAY_DATABASE_MAX_CONNS
|
|
value: "32"
|
|
- name: AI_GATEWAY_DATABASE_CRITICAL_MAX_CONNS
|
|
value: "4"
|
|
- name: AI_GATEWAY_DATABASE_RIVER_MAX_CONNS
|
|
value: "8"
|
|
- name: AI_GATEWAY_DATABASE_MIN_IDLE_CONNS
|
|
value: "4"
|
|
- name: AI_GATEWAY_DATABASE_MAX_CONN_IDLE_SECONDS
|
|
value: "30"
|
|
- name: AI_GATEWAY_ASYNC_WORKER_INSTANCE_HARD_LIMIT
|
|
value: "24"
|
|
- name: AI_GATEWAY_ASYNC_ADMISSION_MICROBATCH_SIZE
|
|
value: "8"
|
|
- name: AI_GATEWAY_MEDIA_MATERIALIZATION_CONCURRENCY
|
|
value: "24"
|
|
- name: AI_GATEWAY_MEDIA_REQUEST_CONCURRENCY
|
|
value: "24"
|
|
- name: AI_GATEWAY_MEDIA_IMAGE_NORMALIZATION_CONCURRENCY
|
|
value: "2"
|
|
- name: AI_GATEWAY_MEDIA_OSS_ACCESS_KEY_ID
|
|
valueFrom:
|
|
secretKeyRef:
|
|
name: easyai-oss-backup
|
|
key: ACCESS_KEY_ID
|
|
- name: AI_GATEWAY_MEDIA_OSS_ACCESS_KEY_SECRET
|
|
valueFrom:
|
|
secretKeyRef:
|
|
name: easyai-oss-backup
|
|
key: ACCESS_SECRET_KEY
|
|
- name: POD_NAMESPACE
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: metadata.namespace
|
|
- name: POD_NAME
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: metadata.name
|
|
- name: POD_UID
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: metadata.uid
|
|
- name: POD_IP
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: status.podIP
|
|
- name: AI_GATEWAY_WORKER_ID
|
|
value: pool-hk-1-worker
|
|
- name: AI_GATEWAY_WORKER_ADVERTISE_ENDPOINT
|
|
value: http://$(POD_IP):8088
|
|
- name: AI_GATEWAY_ORCHESTRATOR_INSTANCE_REF
|
|
valueFrom:
|
|
fieldRef:
|
|
fieldPath: metadata.name
|
|
- name: AI_GATEWAY_EXECUTION_POOL_ID
|
|
value: pool-hk-1
|
|
- name: AI_GATEWAY_EXECUTION_POOL_LABELS
|
|
value: '{"region":"ap-east","zone":"hongkong","provider":"self-hosted","egress_profile":"direct"}'
|
|
- name: AI_GATEWAY_PLATFORM_PROXY_BYPASS_IDS
|
|
value: 99372d7c-f2a4-472a-987f-30cb76c7962c
|
|
ports:
|
|
- name: health
|
|
containerPort: 8088
|
|
startupProbe:
|
|
httpGet:
|
|
path: /healthz
|
|
port: health
|
|
periodSeconds: 5
|
|
timeoutSeconds: 3
|
|
failureThreshold: 120
|
|
readinessProbe:
|
|
httpGet:
|
|
path: /readyz
|
|
port: health
|
|
initialDelaySeconds: 3
|
|
periodSeconds: 5
|
|
timeoutSeconds: 3
|
|
failureThreshold: 6
|
|
livenessProbe:
|
|
httpGet:
|
|
path: /healthz
|
|
port: health
|
|
initialDelaySeconds: 15
|
|
periodSeconds: 10
|
|
timeoutSeconds: 3
|
|
failureThreshold: 6
|
|
lifecycle:
|
|
preStop:
|
|
exec:
|
|
command: ["/bin/sh", "-c", "sleep 10"]
|
|
resources:
|
|
requests:
|
|
cpu: 500m
|
|
memory: 1536Mi
|
|
limits:
|
|
cpu: "2"
|
|
memory: 2Gi
|
|
securityContext:
|
|
allowPrivilegeEscalation: false
|
|
capabilities:
|
|
drop: ["ALL"]
|
|
readOnlyRootFilesystem: true
|
|
volumeMounts:
|
|
- name: application-data
|
|
mountPath: /app/data
|
|
- name: temporary-files
|
|
mountPath: /tmp
|
|
volumes:
|
|
- name: application-data
|
|
emptyDir:
|
|
sizeLimit: 5Gi
|
|
- name: temporary-files
|
|
emptyDir:
|
|
sizeLimit: 1Gi
|
|
---
|
|
apiVersion: apps/v1
|
|
kind: Deployment
|
|
metadata:
|
|
name: easyai-web-ningbo
|
|
labels:
|
|
app.kubernetes.io/name: easyai-web
|
|
app.kubernetes.io/component: web
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
spec:
|
|
replicas: 1
|
|
strategy:
|
|
type: RollingUpdate
|
|
rollingUpdate:
|
|
maxSurge: 1
|
|
maxUnavailable: 0
|
|
selector:
|
|
matchLabels:
|
|
app.kubernetes.io/name: easyai-web
|
|
easyai.io/site: ningbo
|
|
template:
|
|
metadata:
|
|
labels:
|
|
app.kubernetes.io/name: easyai-web
|
|
app.kubernetes.io/component: web
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
easyai.io/site: ningbo
|
|
spec:
|
|
automountServiceAccountToken: false
|
|
terminationGracePeriodSeconds: 30
|
|
nodeSelector:
|
|
easyai.io/site: ningbo
|
|
easyai.io/workload: "true"
|
|
securityContext:
|
|
runAsNonRoot: true
|
|
runAsUser: 101
|
|
runAsGroup: 101
|
|
fsGroup: 101
|
|
fsGroupChangePolicy: OnRootMismatch
|
|
seccompProfile:
|
|
type: RuntimeDefault
|
|
containers:
|
|
- name: web
|
|
image: easyai-web
|
|
imagePullPolicy: IfNotPresent
|
|
ports:
|
|
- name: http
|
|
containerPort: 80
|
|
readinessProbe:
|
|
httpGet:
|
|
path: /
|
|
port: http
|
|
initialDelaySeconds: 3
|
|
periodSeconds: 5
|
|
timeoutSeconds: 3
|
|
failureThreshold: 6
|
|
livenessProbe:
|
|
httpGet:
|
|
path: /
|
|
port: http
|
|
initialDelaySeconds: 15
|
|
periodSeconds: 10
|
|
timeoutSeconds: 3
|
|
failureThreshold: 6
|
|
resources:
|
|
requests:
|
|
cpu: 50m
|
|
memory: 64Mi
|
|
limits:
|
|
cpu: 500m
|
|
memory: 256Mi
|
|
securityContext:
|
|
allowPrivilegeEscalation: false
|
|
capabilities:
|
|
drop: ["ALL"]
|
|
add: ["NET_BIND_SERVICE"]
|
|
readOnlyRootFilesystem: true
|
|
volumeMounts:
|
|
- name: nginx-cache
|
|
mountPath: /var/cache/nginx
|
|
- name: nginx-run
|
|
mountPath: /var/run
|
|
volumes:
|
|
- name: nginx-cache
|
|
emptyDir:
|
|
sizeLimit: 128Mi
|
|
- name: nginx-run
|
|
emptyDir:
|
|
sizeLimit: 16Mi
|
|
---
|
|
apiVersion: apps/v1
|
|
kind: Deployment
|
|
metadata:
|
|
name: easyai-web-hongkong
|
|
labels:
|
|
app.kubernetes.io/name: easyai-web
|
|
app.kubernetes.io/component: web
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
spec:
|
|
replicas: 1
|
|
strategy:
|
|
type: RollingUpdate
|
|
rollingUpdate:
|
|
maxSurge: 1
|
|
maxUnavailable: 0
|
|
selector:
|
|
matchLabels:
|
|
app.kubernetes.io/name: easyai-web
|
|
easyai.io/site: hongkong
|
|
template:
|
|
metadata:
|
|
labels:
|
|
app.kubernetes.io/name: easyai-web
|
|
app.kubernetes.io/component: web
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
easyai.io/site: hongkong
|
|
spec:
|
|
automountServiceAccountToken: false
|
|
terminationGracePeriodSeconds: 30
|
|
nodeSelector:
|
|
easyai.io/site: hongkong
|
|
easyai.io/workload: "true"
|
|
securityContext:
|
|
runAsNonRoot: true
|
|
runAsUser: 101
|
|
runAsGroup: 101
|
|
fsGroup: 101
|
|
fsGroupChangePolicy: OnRootMismatch
|
|
seccompProfile:
|
|
type: RuntimeDefault
|
|
containers:
|
|
- name: web
|
|
image: easyai-web
|
|
imagePullPolicy: IfNotPresent
|
|
ports:
|
|
- name: http
|
|
containerPort: 80
|
|
readinessProbe:
|
|
httpGet:
|
|
path: /
|
|
port: http
|
|
initialDelaySeconds: 3
|
|
periodSeconds: 5
|
|
timeoutSeconds: 3
|
|
failureThreshold: 6
|
|
livenessProbe:
|
|
httpGet:
|
|
path: /
|
|
port: http
|
|
initialDelaySeconds: 15
|
|
periodSeconds: 10
|
|
timeoutSeconds: 3
|
|
failureThreshold: 6
|
|
resources:
|
|
requests:
|
|
cpu: 50m
|
|
memory: 64Mi
|
|
limits:
|
|
cpu: 500m
|
|
memory: 256Mi
|
|
securityContext:
|
|
allowPrivilegeEscalation: false
|
|
capabilities:
|
|
drop: ["ALL"]
|
|
add: ["NET_BIND_SERVICE"]
|
|
readOnlyRootFilesystem: true
|
|
volumeMounts:
|
|
- name: nginx-cache
|
|
mountPath: /var/cache/nginx
|
|
- name: nginx-run
|
|
mountPath: /var/run
|
|
volumes:
|
|
- name: nginx-cache
|
|
emptyDir:
|
|
sizeLimit: 128Mi
|
|
- name: nginx-run
|
|
emptyDir:
|
|
sizeLimit: 16Mi
|
|
---
|
|
apiVersion: v1
|
|
kind: Service
|
|
metadata:
|
|
name: api
|
|
labels:
|
|
app.kubernetes.io/name: easyai-api
|
|
app.kubernetes.io/component: api
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
spec:
|
|
type: NodePort
|
|
externalTrafficPolicy: Local
|
|
selector:
|
|
app.kubernetes.io/name: easyai-api
|
|
ports:
|
|
- name: http
|
|
port: 8088
|
|
targetPort: http
|
|
nodePort: 30088
|
|
---
|
|
apiVersion: v1
|
|
kind: Service
|
|
metadata:
|
|
name: web
|
|
labels:
|
|
app.kubernetes.io/name: easyai-web
|
|
app.kubernetes.io/component: web
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
spec:
|
|
type: NodePort
|
|
externalTrafficPolicy: Local
|
|
selector:
|
|
app.kubernetes.io/name: easyai-web
|
|
ports:
|
|
- name: http
|
|
port: 80
|
|
targetPort: http
|
|
nodePort: 30178
|
|
---
|
|
apiVersion: v1
|
|
kind: Service
|
|
metadata:
|
|
name: api-ningbo-edge
|
|
labels:
|
|
app.kubernetes.io/name: easyai-api
|
|
app.kubernetes.io/component: api
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
easyai.io/site: ningbo
|
|
spec:
|
|
type: NodePort
|
|
externalTrafficPolicy: Local
|
|
selector:
|
|
app.kubernetes.io/name: easyai-api
|
|
easyai.io/site: ningbo
|
|
ports:
|
|
- name: http
|
|
port: 8088
|
|
targetPort: http
|
|
nodePort: 31088
|
|
---
|
|
apiVersion: v1
|
|
kind: Service
|
|
metadata:
|
|
name: api-hongkong-edge
|
|
labels:
|
|
app.kubernetes.io/name: easyai-api
|
|
app.kubernetes.io/component: api
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
easyai.io/site: hongkong
|
|
spec:
|
|
type: NodePort
|
|
externalTrafficPolicy: Local
|
|
selector:
|
|
app.kubernetes.io/name: easyai-api
|
|
easyai.io/site: hongkong
|
|
ports:
|
|
- name: http
|
|
port: 8088
|
|
targetPort: http
|
|
nodePort: 31089
|
|
---
|
|
apiVersion: v1
|
|
kind: Service
|
|
metadata:
|
|
name: web-ningbo-edge
|
|
labels:
|
|
app.kubernetes.io/name: easyai-web
|
|
app.kubernetes.io/component: web
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
easyai.io/site: ningbo
|
|
spec:
|
|
type: NodePort
|
|
externalTrafficPolicy: Local
|
|
selector:
|
|
app.kubernetes.io/name: easyai-web
|
|
easyai.io/site: ningbo
|
|
ports:
|
|
- name: http
|
|
port: 80
|
|
targetPort: http
|
|
nodePort: 31178
|
|
---
|
|
apiVersion: v1
|
|
kind: Service
|
|
metadata:
|
|
name: web-hongkong-edge
|
|
labels:
|
|
app.kubernetes.io/name: easyai-web
|
|
app.kubernetes.io/component: web
|
|
app.kubernetes.io/part-of: easyai-ai-gateway
|
|
easyai.io/site: hongkong
|
|
spec:
|
|
type: NodePort
|
|
externalTrafficPolicy: Local
|
|
selector:
|
|
app.kubernetes.io/name: easyai-web
|
|
easyai.io/site: hongkong
|
|
ports:
|
|
- name: http
|
|
port: 80
|
|
targetPort: http
|
|
nodePort: 31179
|
|
---
|
|
apiVersion: policy/v1
|
|
kind: PodDisruptionBudget
|
|
metadata:
|
|
name: easyai-api
|
|
spec:
|
|
minAvailable: 1
|
|
selector:
|
|
matchLabels:
|
|
app.kubernetes.io/name: easyai-api
|
|
---
|
|
apiVersion: policy/v1
|
|
kind: PodDisruptionBudget
|
|
metadata:
|
|
name: easyai-web
|
|
spec:
|
|
minAvailable: 1
|
|
selector:
|
|
matchLabels:
|
|
app.kubernetes.io/name: easyai-web
|
|
---
|
|
apiVersion: policy/v1
|
|
kind: PodDisruptionBudget
|
|
metadata:
|
|
name: easyai-capacity-controller
|
|
spec:
|
|
minAvailable: 1
|
|
selector:
|
|
matchLabels:
|
|
app.kubernetes.io/name: easyai-capacity-controller
|
|
---
|
|
apiVersion: policy/v1
|
|
kind: PodDisruptionBudget
|
|
metadata:
|
|
name: easyai-worker
|
|
spec:
|
|
minAvailable: 1
|
|
selector:
|
|
matchLabels:
|
|
app.kubernetes.io/name: easyai-worker
|