easyai-ai-gateway/deploy/kubernetes/production/database.yaml

apiVersion: v1
kind: ConfigMap
metadata:
  name: easyai-barman-aws-config
  labels:
    app.kubernetes.io/part-of: easyai-ai-gateway
data:
  config: |
    [default]
    region = oss-cn-shanghai
    s3 =
      addressing_style = virtual
      signature_version = s3
---
apiVersion: barmancloud.cnpg.io/v1
kind: ObjectStore
metadata:
  name: easyai-postgres-oss
  labels:
    app.kubernetes.io/part-of: easyai-ai-gateway
spec:
  retentionPolicy: 14d
  configuration:
    destinationPath: s3://wangbo0808/easyai-ai-gateway/production/postgresql
    endpointURL: https://s3.oss-cn-shanghai.aliyuncs.com
    s3Credentials:
      accessKeyId:
        name: easyai-oss-backup
        key: ACCESS_KEY_ID
      secretAccessKey:
        name: easyai-oss-backup
        key: ACCESS_SECRET_KEY
    wal:
      archiveAdditionalCommandArgs:
        - --addressing-style=virtual
      compression: gzip
      maxParallel: 2
      restoreAdditionalCommandArgs:
        - --addressing-style=virtual
    data:
      additionalCommandArgs:
        - --addressing-style=virtual
      compression: gzip
      jobs: 2
      restoreAdditionalCommandArgs:
        - --addressing-style=virtual
  instanceSidecarConfiguration:
    env:
      - name: AWS_CONFIG_FILE
        value: /projected/barman-aws/config
      - name: AWS_DEFAULT_REGION
        value: oss-cn-shanghai
      - name: AWS_REQUEST_CHECKSUM_CALCULATION
        value: when_required
      - name: AWS_RESPONSE_CHECKSUM_VALIDATION
        value: when_required
---
apiVersion: postgresql.cnpg.io/v1
kind: Cluster
metadata:
  name: easyai-postgres
  labels:
    app.kubernetes.io/name: easyai-postgres
    app.kubernetes.io/part-of: easyai-ai-gateway
spec:
  # Production currently uses one Ningbo primary. Cross-region replicas are
  # intentionally disabled until a separately certified HA topology exists.
  instances: 1
  imageName: ghcr.io/cloudnative-pg/postgresql:18.4-standard-trixie@sha256:4587df73024408f5b2be9b4dd6ba2ccee8c9e5dc0c9a87c274c292291cc8a68c
  imagePullPolicy: IfNotPresent
  env:
    - name: KUBERNETES_SERVICE_HOST
      valueFrom:
        fieldRef:
          fieldPath: spec.nodeName
    - name: KUBERNETES_SERVICE_PORT
      value: "6443"
  primaryUpdateStrategy: unsupervised
  primaryUpdateMethod: switchover
  failoverDelay: 0
  switchoverDelay: 60
  stopDelay: 90
  smartShutdownTimeout: 30
  livenessProbeTimeout: 300
  probes:
    liveness:
      isolationCheck:
        enabled: true
        connectionTimeout: 5000
        requestTimeout: 5000
  bootstrap:
    initdb:
      database: easyai_ai_gateway
      owner: easyai
      secret:
        name: easyai-postgres-app
      postInitSQL:
        - CREATE EXTENSION IF NOT EXISTS pgcrypto;
  storage:
    storageClass: local-path
    size: 20Gi
  walStorage:
    storageClass: local-path
    size: 5Gi
  affinity:
    enablePodAntiAffinity: true
    podAntiAffinityType: required
    topologyKey: easyai.io/site
    nodeSelector:
      easyai.io/database: "true"
      easyai.io/site: ningbo
  resources:
    requests:
      cpu: 500m
      memory: 1Gi
    limits:
      cpu: "2"
      memory: 3Gi
  postgresql:
    # Keep this explicit null so server-side apply atomically removes the old
    # two-instance synchronous setting during preflight field management.
    synchronous: null
    parameters:
      # A 16 MiB WAL segment every 60 seconds requires at least 2.24 Mbit/s
      # of continuous cross-region throughput even while the database is idle.
      # The certified Ningbo/Hong Kong path is below that floor, so keep the
      # object-storage RPO within five minutes without starving sync replay.
      archive_timeout: 5min
      max_connections: "200"
      shared_buffers: 512MB
      wal_compression: "on"
      wal_keep_size: 1GB
      # CNPG defaults these cross-region liveness timers to 5 seconds. A short
      # control-plane or pod-network stall can otherwise tear down a future
      # WAL stream. Keep failure detection bounded without amplifying a
      # transient multi-second stall into repeated reconnects.
      wal_sender_timeout: 30s
      wal_receiver_timeout: 30s
  monitoring:
    enablePodMonitor: false
  projectedVolumeTemplate:
    defaultMode: 0444
    sources:
      - configMap:
          name: easyai-barman-aws-config
          items:
            - key: config
              path: barman-aws/config
  plugins:
    - name: barman-cloud.cloudnative-pg.io
      isWALArchiver: true
      parameters:
        barmanObjectName: easyai-postgres-oss
---
apiVersion: postgresql.cnpg.io/v1
kind: ScheduledBackup
metadata:
  name: easyai-postgres-daily
  labels:
    app.kubernetes.io/part-of: easyai-ai-gateway
spec:
  # Six-field cron, UTC: 18:30 UTC is 02:30 Asia/Shanghai.
  schedule: "0 30 18 * * *"
  backupOwnerReference: self
  cluster:
    name: easyai-postgres
  method: plugin
  pluginConfiguration:
    name: barman-cloud.cloudnative-pg.io
  immediate: false
  suspend: false