chore: add imagePullPolicy and migrate WaitGroup to wg.Go

- add imagePullPolicy: IfNotPresent to all k8s Deployments, DaemonSet
    (grafana, jaeger, loki, rabbitmq, redis, promtail)
  - migrate wg.Add(1)/go/defer wg.Done() pattern to wg.Go() (Go 1.25+)
    in logger/loki_syncer.go and task/worker.go
  - simplify redundant map existence check before delete in diagram/graph.go
  - update deploy.md to reflect pg PVC size (6Gi) and resource limits

.gitignore

@@ -32,6 +32,7 @@ go.work
 # ai config
 .cursor/
 .claude/
+.codewhale/
 .cursorrules
 .copilot/
 .chatgpt/

deploy/deploy.md

@@ -695,7 +695,11 @@ kubectl apply -f deploy/k8s/pg-service.yaml
 | **数据库** | `demo` | ConfigMap 中 `POSTGRES_DB` |
 | **用户名** | `postgres` | ConfigMap 中 `POSTGRES_USER` |
 | **密码** | `coslight` | ConfigMap `postgres-config` 中配置，生产环境迁移至 Secret |
-| **存储** | `2Gi` | PVC `postgres-data` |
+| **存储** | `6Gi` | PVC `postgres-data` |
+| **CPU** | `100m` 请求 / `500m` 上限 | StatefulSet `resources` 字段 |
+| **内存** | `256Mi` 请求 / `512Mi` 上限 | StatefulSet `resources` 字段 |
+
+> **注意：** 密码当前以明文形式存储在 `pg-configmap.yaml` 中，生产环境应将其迁移至 K8s Secret，并通过环境变量注入容器，避免将明文密码提交至版本库。
 
 ##### 4.4.1 等待 Pod 就绪
 

deploy/k8s/grafana-deployment.yaml

@@ -16,6 +16,7 @@ spec:
       containers:
         - name: grafana
           image: grafana/grafana:10.4.2
+          imagePullPolicy: IfNotPresent
           ports:
             - containerPort: 3000
           env:

deploy/k8s/jaeger-deployment.yaml

@@ -15,6 +15,7 @@ spec:
       containers:
         - name: jaeger
           image: jaegertracing/all-in-one:1.56
+          imagePullPolicy: IfNotPresent
           env:
             - name: COLLECTOR_OTLP_ENABLED
               value: "true"

deploy/k8s/loki-deployment.yaml

@@ -20,6 +20,7 @@ spec:
       containers:
         - name: loki
           image: grafana/loki:2.9.4
+          imagePullPolicy: IfNotPresent
           args:
             - -config.file=/etc/loki/loki.yaml
           ports:

deploy/k8s/mongodb-statefulset.yaml

@@ -34,9 +34,9 @@ spec:
                 - mongosh
                 - --eval
                 - "db.adminCommand('ping')"
-            initialDelaySeconds: 10
+            initialDelaySeconds: 30
-            periodSeconds: 5
+            periodSeconds: 10
-            timeoutSeconds: 3
+            timeoutSeconds: 10
             failureThreshold: 12
           livenessProbe:
             exec:
@@ -44,10 +44,10 @@ spec:
                 - mongosh
                 - --eval
                 - "db.adminCommand('ping')"
-            initialDelaySeconds: 30
+            initialDelaySeconds: 120
-            periodSeconds: 20
+            periodSeconds: 10
-            timeoutSeconds: 3
+            timeoutSeconds: 30
-            failureThreshold: 3
+            failureThreshold: 5
           resources:
             requests:
               cpu: 100m

deploy/k8s/promtail-daemonset.yaml

@@ -19,6 +19,7 @@ spec:
       containers:
         - name: promtail
           image: grafana/promtail:2.9.4
+          imagePullPolicy: IfNotPresent
           args:
             - -config.file=/etc/promtail/promtail.yaml
           ports:

deploy/k8s/rabbitmq-deployment.yaml

@@ -15,6 +15,7 @@ spec:
       containers:
         - name: rabbitmq
           image: rabbitmq:4.1.1-management-alpine
+          imagePullPolicy: IfNotPresent
           ports:
             - containerPort: 4369
             - containerPort: 5671

deploy/k8s/redis-deployment.yaml

@@ -15,6 +15,7 @@ spec:
       containers:
         - name: redis
           image: redis/redis-stack-server:latest
+          imagePullPolicy: IfNotPresent
           resources:
             limits:
               memory: "128Mi"

diagram/graph.go

@@ -65,9 +65,7 @@ func (g *Graph) AddEdge(from, to uuid.UUID) {
 
 	// 创建新的拓扑信息时，如果被链接的点已经存在于游离节点中
 	// 则将其移除
-	if _, exist := g.FreeVertexs[toKey]; exist {
+	delete(g.FreeVertexs, toKey)
-		delete(g.FreeVertexs, toKey)
-	}
 }
 
 // DelNode delete a node to the graph

logger/loki_syncer.go

@@ -47,8 +47,7 @@ func newLokiSyncer(lCfg config.LokiConfig) *lokiSyncer {
 		client:   &http.Client{Timeout: 5 * time.Second},
 		ch:       make(chan string, 512),
 	}
-	ls.wg.Add(1)
+	ls.wg.Go(ls.run)
-	go ls.run()
 	return ls
 }
 
@@ -70,7 +69,6 @@ func (ls *lokiSyncer) Sync() error {
 }
 
 func (ls *lokiSyncer) run() {
-	defer ls.wg.Done()
 	ticker := time.NewTicker(2 * time.Second)
 	defer ticker.Stop()
 

task/worker.go

@@ -185,13 +185,11 @@ func (w *TaskWorker) Start() error {
 
 	// Start multiple consumers for better throughput
 	for i := 0; i < w.cfg.QueueConsumerCount; i++ {
-		w.wg.Add(1)
+		w.wg.Go(func() { w.consumerLoop(i) })
-		go w.consumerLoop(i)
 	}
 
 	// Start health check goroutine
-	w.wg.Add(1)
+	w.wg.Go(w.healthCheckLoop)
-	go w.healthCheckLoop()
 
 	logger.Info(w.ctx, "task worker started successfully")
 	return nil
@@ -199,8 +197,6 @@ func (w *TaskWorker) Start() error {
 
 // consumerLoop runs a single RabbitMQ consumer
 func (w *TaskWorker) consumerLoop(consumerID int) {
-	defer w.wg.Done()
-
 	logger.Info(w.ctx, "starting consumer", "consumer_id", consumerID)
 
 	// Consume messages from the queue
@@ -478,8 +474,6 @@ func (w *TaskWorker) dispatch(ctx context.Context, taskType TaskType, taskID uui
 
 // healthCheckLoop periodically checks worker health and metrics
 func (w *TaskWorker) healthCheckLoop() {
-	defer w.wg.Done()
-
 	ticker := time.NewTicker(w.cfg.PollingInterval)
 	defer ticker.Stop()