feat(deploy): 切换宁波单节点生产拓扑

移除已停用的香港 API、Web、Worker 和边缘 Service 渲染结果,将 Worker 与容量控制器固定到宁波单实例,并在发布成功后清理旧香港资源。\n\n容量控制器忽略最大副本为零的站点,集群环境允许不再声明香港主机;单节点 production acceptance 不再依赖香港 Service、Deployment 或 SSH 回退。\n\n验证:Go 全量测试与 go vet;迁移安全检查;Kustomize 单节点渲染;ShellCheck;cluster release、production acceptance 和 manual release 测试。
This commit is contained in:
2026-08-03 09:59:34 +08:00
parent 5e9a7888ae
commit 4ff85d96e0
12 changed files with 156 additions and 125 deletions
@@ -46,12 +46,9 @@ data:
AI_GATEWAY_ASYNC_WORKER_GLOBAL_HARD_LIMIT: "48"
AI_GATEWAY_ASYNC_WORKER_REFRESH_INTERVAL_SECONDS: "5"
AI_GATEWAY_WORKER_AUTOSCALING_ENABLED: "false"
AI_GATEWAY_WORKER_REPLICAS_NINGBO: "0"
AI_GATEWAY_WORKER_REPLICAS_HONGKONG: "2"
AI_GATEWAY_WORKER_MIN_REPLICAS_NINGBO: "0"
AI_GATEWAY_WORKER_MIN_REPLICAS_HONGKONG: "1"
AI_GATEWAY_WORKER_MAX_REPLICAS_NINGBO: "0"
AI_GATEWAY_WORKER_MAX_REPLICAS_HONGKONG: "2"
AI_GATEWAY_WORKER_REPLICAS_NINGBO: "1"
AI_GATEWAY_WORKER_MIN_REPLICAS_NINGBO: "1"
AI_GATEWAY_WORKER_MAX_REPLICAS_NINGBO: "1"
AI_GATEWAY_WORKER_TARGET_OUTSTANDING_PER_REPLICA: "48"
AI_GATEWAY_WORKER_SCALE_UP_WINDOW_SECONDS: "20"
AI_GATEWAY_WORKER_SCALE_DOWN_STABILIZATION_SECONDS: "600"