refactor: udpate workflow (#3028)

Signed-off-by: promalert <promalert@outlook.com>

alert/dispatch/dispatch.go

@@ -16,6 +16,7 @@ import (
 	"github.com/ccfos/nightingale/v6/alert/astats"
 	"github.com/ccfos/nightingale/v6/alert/common"
 	"github.com/ccfos/nightingale/v6/alert/pipeline"
+	"github.com/ccfos/nightingale/v6/alert/pipeline/engine"
 	"github.com/ccfos/nightingale/v6/alert/sender"
 	"github.com/ccfos/nightingale/v6/memsto"
 	"github.com/ccfos/nightingale/v6/models"
@@ -231,6 +232,8 @@ func shouldSkipNotify(ctx *ctx.Context, event *models.AlertCurEvent, notifyRuleI
 }
 
 func HandleEventPipeline(pipelineConfigs []models.PipelineConfig, eventOrigin, event *models.AlertCurEvent, eventProcessorCache *memsto.EventProcessorCacheType, ctx *ctx.Context, id int64, from string) *models.AlertCurEvent {
+	workflowEngine := engine.NewWorkflowEngine(ctx)
+
 	for _, pipelineConfig := range pipelineConfigs {
 		if !pipelineConfig.Enable {
 			continue
@@ -247,23 +250,28 @@ func HandleEventPipeline(pipelineConfigs []models.PipelineConfig, eventOrigin, e
 			continue
 		}
 
-		processors := eventProcessorCache.GetProcessorsById(pipelineConfig.PipelineId)
-		for _, processor := range processors {
-			var res string
-			var err error
-			logger.Infof("processor_by_%s_id:%d pipeline_id:%d, before processor:%+v, event: %+v", from, id, pipelineConfig.PipelineId, processor, event)
-			event, res, err = processor.Process(ctx, event)
-			if event == nil {
-				logger.Infof("processor_by_%s_id:%d pipeline_id:%d, event dropped, after processor:%+v, event: %+v", from, id, pipelineConfig.PipelineId, processor, eventOrigin)
-
-				if from == "notify_rule" {
-					// alert_rule 获取不到 eventId 记录没有意义
-					sender.NotifyRecord(ctx, []*models.AlertCurEvent{eventOrigin}, id, "", "", res, fmt.Errorf("processor_by_%s_id:%d pipeline_id:%d, drop by processor", from, id, pipelineConfig.PipelineId))
-				}
-				return nil
-			}
-			logger.Infof("processor_by_%s_id:%d pipeline_id:%d, after processor:%+v, event: %+v, res:%v, err:%v", from, id, pipelineConfig.PipelineId, processor, event, res, err)
+		// 统一使用工作流引擎执行（兼容线性模式和工作流模式）
+		triggerCtx := &models.WorkflowTriggerContext{
+			Mode:      models.TriggerModeEvent,
+			TriggerBy: from,
 		}
+
+		resultEvent, result, err := workflowEngine.Execute(eventPipeline, event, triggerCtx)
+		if err != nil {
+			logger.Errorf("processor_by_%s_id:%d pipeline_id:%d, pipeline execute error: %v", from, id, pipelineConfig.PipelineId, err)
+			continue
+		}
+
+		if resultEvent == nil {
+			logger.Infof("processor_by_%s_id:%d pipeline_id:%d, event dropped, event: %+v", from, id, pipelineConfig.PipelineId, eventOrigin)
+			if from == "notify_rule" {
+				sender.NotifyRecord(ctx, []*models.AlertCurEvent{eventOrigin}, id, "", "", result.Message, fmt.Errorf("processor_by_%s_id:%d pipeline_id:%d, drop by pipeline", from, id, pipelineConfig.PipelineId))
+			}
+			return nil
+		}
+
+		event = resultEvent
+		logger.Infof("processor_by_%s_id:%d pipeline_id:%d, pipeline executed, status:%s, message:%s", from, id, pipelineConfig.PipelineId, result.Status, result.Message)
 	}
 
 	event.FE2DB()
@@ -538,7 +546,7 @@ func SendNotifyRuleMessage(ctx *ctx.Context, userCache *memsto.UserCacheType, us
 		for i := range flashDutyChannelIDs {
 			start := time.Now()
 			respBody, err := notifyChannel.SendFlashDuty(events, flashDutyChannelIDs[i], notifyChannelCache.GetHttpClient(notifyChannel.ID))
-			respBody = fmt.Sprintf("duration: %d ms %s", time.Since(start).Milliseconds(), respBody)
+			respBody = fmt.Sprintf("send_time: %s duration: %d ms %s", time.Now().Format("2006-01-02 15:04:05"), time.Since(start).Milliseconds(), respBody)
 			logger.Infof("duty_sender notify_id: %d, channel_name: %v, event:%+v, IntegrationUrl: %v dutychannel_id: %v, respBody: %v, err: %v", notifyRuleId, notifyChannel.Name, events[0], notifyChannel.RequestConfig.FlashDutyRequestConfig.IntegrationUrl, flashDutyChannelIDs[i], respBody, err)
 			sender.NotifyRecord(ctx, events, notifyRuleId, notifyChannel.Name, strconv.FormatInt(flashDutyChannelIDs[i], 10), respBody, err)
 		}
@@ -547,7 +555,7 @@ func SendNotifyRuleMessage(ctx *ctx.Context, userCache *memsto.UserCacheType, us
 		for _, routingKey := range pagerdutyRoutingKeys {
 			start := time.Now()
 			respBody, err := notifyChannel.SendPagerDuty(events, routingKey, siteInfo.SiteUrl, notifyChannelCache.GetHttpClient(notifyChannel.ID))
-			respBody = fmt.Sprintf("duration: %d ms %s", time.Since(start).Milliseconds(), respBody)
+			respBody = fmt.Sprintf("send_time: %s duration: %d ms %s", time.Now().Format("2006-01-02 15:04:05"), time.Since(start).Milliseconds(), respBody)
 			logger.Infof("pagerduty_sender notify_id: %d, channel_name: %v, event:%+v, respBody: %v, err: %v", notifyRuleId, notifyChannel.Name, events[0], respBody, err)
 			sender.NotifyRecord(ctx, events, notifyRuleId, notifyChannel.Name, "", respBody, err)
 		}
@@ -578,7 +586,7 @@ func SendNotifyRuleMessage(ctx *ctx.Context, userCache *memsto.UserCacheType, us
 	case "script":
 		start := time.Now()
 		target, res, err := notifyChannel.SendScript(events, tplContent, customParams, sendtos)
-		res = fmt.Sprintf("duration: %d ms %s", time.Since(start).Milliseconds(), res)
+		res = fmt.Sprintf("send_time: %s duration: %d ms %s", time.Now().Format("2006-01-02 15:04:05"), time.Since(start).Milliseconds(), res)
 		logger.Infof("script_sender notify_id: %d, channel_name: %v, event:%+v, tplContent:%s, customParams:%v, target:%s, res:%s, err:%v", notifyRuleId, notifyChannel.Name, events[0], tplContent, customParams, target, res, err)
 		sender.NotifyRecord(ctx, events, notifyRuleId, notifyChannel.Name, target, res, err)
 	default:

alert/eval/eval.go

@@ -11,6 +11,7 @@ import (
 	"strconv"
 	"strings"
 	"sync"
+	"text/template"
 	"time"
 
 	"github.com/ccfos/nightingale/v6/alert/astats"
@@ -24,6 +25,7 @@ import (
 	"github.com/ccfos/nightingale/v6/pkg/poster"
 	promsdk "github.com/ccfos/nightingale/v6/pkg/prom"
 	promql2 "github.com/ccfos/nightingale/v6/pkg/promql"
+	"github.com/ccfos/nightingale/v6/pkg/tplx"
 	"github.com/ccfos/nightingale/v6/pkg/unit"
 	"github.com/ccfos/nightingale/v6/prom"
 	"github.com/prometheus/common/model"
@@ -60,6 +62,7 @@ const (
 	CHECK_QUERY     = "check_query_config"
 	GET_CLIENT      = "get_client"
 	QUERY_DATA      = "query_data"
+	EXEC_TEMPLATE   = "exec_template"
 )
 
 const (
@@ -1483,6 +1486,16 @@ func (arw *AlertRuleWorker) GetAnomalyPoint(rule *models.AlertRule, dsId int64)
 				return points, recoverPoints, fmt.Errorf("rule_eval:%d datasource:%d not exists", rule.Id, dsId)
 			}
 
+			if err = ExecuteQueryTemplate(rule.Cate, query, nil); err != nil {
+				logger.Warningf("rule_eval rid:%d execute query template error: %v", rule.Id, err)
+				arw.Processor.Stats.CounterRuleEvalErrorTotal.WithLabelValues(fmt.Sprintf("%v", arw.Processor.DatasourceId()), EXEC_TEMPLATE, arw.Processor.BusiGroupCache.GetNameByBusiGroupId(arw.Rule.GroupId), fmt.Sprintf("%v", arw.Rule.Id)).Inc()
+				arw.Processor.Stats.GaugeQuerySeriesCount.WithLabelValues(
+					fmt.Sprintf("%v", arw.Rule.Id),
+					fmt.Sprintf("%v", arw.Processor.DatasourceId()),
+					fmt.Sprintf("%v", i),
+				).Set(-3)
+			}
+
 			ctx := context.WithValue(context.Background(), "delay", int64(rule.Delay))
 			series, err := plug.QueryData(ctx, query)
 			arw.Processor.Stats.CounterQueryDataTotal.WithLabelValues(fmt.Sprintf("%d", arw.DatasourceId), fmt.Sprintf("%d", rule.Id)).Inc()
@@ -1698,3 +1711,61 @@ func (arw *AlertRuleWorker) GetAnomalyPoint(rule *models.AlertRule, dsId int64)
 
 	return points, recoverPoints, nil
 }
+
+// ExecuteQueryTemplate 根据数据源类型对 Query 进行模板渲染处理
+// cate: 数据源类别，如 "mysql", "pgsql" 等
+// query: 查询对象，如果是数据库类型的数据源，会处理其中的 sql 字段
+// data: 模板数据对象，如果为 nil 则使用空结构体（不支持变量渲染），如果不为 nil 则使用传入的数据（支持变量渲染）
+func ExecuteQueryTemplate(cate string, query interface{}, data interface{}) error {
+	// 检查 query 是否是 map，且包含 sql 字段
+	queryMap, ok := query.(map[string]interface{})
+	if !ok {
+		return nil
+	}
+
+	sqlVal, exists := queryMap["sql"]
+	if !exists {
+		return nil
+	}
+
+	sqlStr, ok := sqlVal.(string)
+	if !ok {
+		return nil
+	}
+
+	// 调用 ExecuteSqlTemplate 处理 sql 字段
+	processedSQL, err := ExecuteSqlTemplate(sqlStr, data)
+	if err != nil {
+		return fmt.Errorf("execute sql template error: %w", err)
+	}
+
+	// 更新 query 中的 sql 字段
+	queryMap["sql"] = processedSQL
+	return nil
+}
+
+// ExecuteSqlTemplate 执行 query 中的 golang 模板语法函数
+// query: 要处理的 query 字符串
+// data: 模板数据对象，如果为 nil 则使用空结构体（不支持变量渲染），如果不为 nil 则使用传入的数据（支持变量渲染）
+func ExecuteSqlTemplate(query string, data interface{}) (string, error) {
+	if !strings.Contains(query, "{{") || !strings.Contains(query, "}}") {
+		return query, nil
+	}
+
+	tmpl, err := template.New("query").Funcs(tplx.TemplateFuncMap).Parse(query)
+	if err != nil {
+		return "", fmt.Errorf("query tmpl parse error: %w", err)
+	}
+
+	var buf strings.Builder
+	templateData := data
+	if templateData == nil {
+		templateData = struct{}{}
+	}
+
+	if err := tmpl.Execute(&buf, templateData); err != nil {
+		return "", fmt.Errorf("query tmpl execute error: %w", err)
+	}
+
+	return buf.String(), nil
+}

alert/pipeline/engine/engine.go

@@ -0,0 +1,383 @@
+package engine
+
+import (
+	"fmt"
+	"time"
+
+	"github.com/ccfos/nightingale/v6/models"
+	"github.com/ccfos/nightingale/v6/pkg/ctx"
+	"github.com/google/uuid"
+	"github.com/toolkits/pkg/logger"
+)
+
+type WorkflowEngine struct {
+	ctx *ctx.Context
+}
+
+func NewWorkflowEngine(c *ctx.Context) *WorkflowEngine {
+	return &WorkflowEngine{ctx: c}
+}
+
+func (e *WorkflowEngine) Execute(pipeline *models.EventPipeline, event *models.AlertCurEvent, triggerCtx *models.WorkflowTriggerContext) (*models.AlertCurEvent, *models.WorkflowResult, error) {
+	startTime := time.Now()
+
+	wfCtx := e.initWorkflowContext(pipeline, event, triggerCtx)
+
+	nodes := pipeline.GetWorkflowNodes()
+	connections := pipeline.GetWorkflowConnections()
+
+	if len(nodes) == 0 {
+		return event, &models.WorkflowResult{
+			Event:   event,
+			Status:  models.ExecutionStatusSuccess,
+			Message: "no nodes to execute",
+		}, nil
+	}
+
+	nodeMap := make(map[string]*models.WorkflowNode)
+	for i := range nodes {
+		if nodes[i].RetryInterval == 0 {
+			nodes[i].RetryInterval = 1
+		}
+
+		if nodes[i].MaxRetries == 0 {
+			nodes[i].MaxRetries = 1
+		}
+
+		nodeMap[nodes[i].ID] = &nodes[i]
+	}
+
+	result := e.executeDAG(nodeMap, connections, wfCtx)
+	result.Event = wfCtx.Event
+
+	duration := time.Since(startTime).Milliseconds()
+
+	if triggerCtx != nil && triggerCtx.Mode != "" {
+		e.saveExecutionRecord(pipeline, wfCtx, result, triggerCtx, startTime.Unix(), duration)
+	}
+
+	return wfCtx.Event, result, nil
+}
+
+func (e *WorkflowEngine) initWorkflowContext(pipeline *models.EventPipeline, event *models.AlertCurEvent, triggerCtx *models.WorkflowTriggerContext) *models.WorkflowContext {
+	// 合并环境变量
+	env := pipeline.GetEnvMap()
+	if triggerCtx != nil && triggerCtx.EnvOverrides != nil {
+		for k, v := range triggerCtx.EnvOverrides {
+			env[k] = v
+		}
+	}
+
+	metadata := map[string]string{
+		"start_time":  fmt.Sprintf("%d", time.Now().Unix()),
+		"pipeline_id": fmt.Sprintf("%d", pipeline.ID),
+	}
+
+	// 是否启用流式输出
+	stream := false
+	if triggerCtx != nil {
+		metadata["request_id"] = triggerCtx.RequestID
+		metadata["trigger_mode"] = triggerCtx.Mode
+		metadata["trigger_by"] = triggerCtx.TriggerBy
+		stream = triggerCtx.Stream
+	}
+
+	return &models.WorkflowContext{
+		Event:    event,
+		Env:      env,
+		Vars:     make(map[string]interface{}), // 初始化空的 Vars，供节点间传递数据
+		Metadata: metadata,
+		Stream:   stream,
+	}
+}
+
+// executeDAG 使用 Kahn 算法执行 DAG
+func (e *WorkflowEngine) executeDAG(nodeMap map[string]*models.WorkflowNode, connections models.Connections, wfCtx *models.WorkflowContext) *models.WorkflowResult {
+	result := &models.WorkflowResult{
+		Status:      models.ExecutionStatusSuccess,
+		NodeResults: make([]*models.NodeExecutionResult, 0),
+		Stream:      wfCtx.Stream, // 从上下文继承流式输出设置
+	}
+
+	// 计算每个节点的入度
+	inDegree := make(map[string]int)
+	for nodeID := range nodeMap {
+		inDegree[nodeID] = 0
+	}
+
+	// 遍历连接，计算入度
+	for _, nodeConns := range connections {
+		for _, targets := range nodeConns.Main {
+			for _, target := range targets {
+				inDegree[target.Node]++
+			}
+		}
+	}
+
+	// 找到所有入度为 0 的节点（起始节点）
+	queue := make([]string, 0)
+	for nodeID, degree := range inDegree {
+		if degree == 0 {
+			queue = append(queue, nodeID)
+		}
+	}
+
+	// 如果没有起始节点，说明存在循环依赖
+	if len(queue) == 0 && len(nodeMap) > 0 {
+		result.Status = models.ExecutionStatusFailed
+		result.Message = "workflow has circular dependency"
+		return result
+	}
+
+	// 记录已执行的节点
+	executed := make(map[string]bool)
+	// 记录节点的分支选择结果
+	branchResults := make(map[string]*int)
+
+	for len(queue) > 0 {
+		// 取出队首节点
+		nodeID := queue[0]
+		queue = queue[1:]
+
+		// 检查是否已执行
+		if executed[nodeID] {
+			continue
+		}
+
+		node, exists := nodeMap[nodeID]
+		if !exists {
+			continue
+		}
+
+		// 执行节点
+		nodeResult, nodeOutput := e.executeNode(node, wfCtx)
+		result.NodeResults = append(result.NodeResults, nodeResult)
+
+		if nodeOutput != nil && nodeOutput.Stream && nodeOutput.StreamChan != nil {
+			// 流式输出节点通常是最后一个节点
+			// 直接传递 StreamChan 给 WorkflowResult，不阻塞等待
+			result.Stream = true
+			result.StreamChan = nodeOutput.StreamChan
+			result.Event = wfCtx.Event
+			result.Status = "streaming"
+			result.Message = fmt.Sprintf("streaming output from node: %s", node.Name)
+
+			// 更新节点状态为 streaming
+			nodeResult.Status = "streaming"
+			nodeResult.Message = "streaming in progress"
+
+			// 立即返回，让 API 层处理流式响应
+			return result
+		}
+		executed[nodeID] = true
+
+		// 保存分支结果
+		if nodeResult.BranchIndex != nil {
+			branchResults[nodeID] = nodeResult.BranchIndex
+		}
+
+		// 检查执行状态
+		if nodeResult.Status == "failed" {
+			if !node.ContinueOnFail {
+				result.Status = models.ExecutionStatusFailed
+				result.ErrorNode = nodeID
+				result.Message = fmt.Sprintf("node %s failed: %s", node.Name, nodeResult.Error)
+				return result
+			}
+		}
+
+		// 检查是否终止
+		if nodeResult.Status == "terminated" {
+			result.Message = fmt.Sprintf("workflow terminated at node %s", node.Name)
+			return result
+		}
+
+		// 更新后继节点的入度
+		if nodeConns, ok := connections[nodeID]; ok {
+			for outputIndex, targets := range nodeConns.Main {
+				// 检查是否应该走这个分支
+				if !e.shouldFollowBranch(nodeID, outputIndex, branchResults) {
+					continue
+				}
+
+				for _, target := range targets {
+					inDegree[target.Node]--
+					if inDegree[target.Node] == 0 {
+						queue = append(queue, target.Node)
+					}
+				}
+			}
+		}
+	}
+
+	return result
+}
+
+// executeNode 执行单个节点
+// 返回：节点执行结果、节点输出（用于流式输出检测）
+func (e *WorkflowEngine) executeNode(node *models.WorkflowNode, wfCtx *models.WorkflowContext) (*models.NodeExecutionResult, *models.NodeOutput) {
+	startTime := time.Now()
+	nodeResult := &models.NodeExecutionResult{
+		NodeID:    node.ID,
+		NodeName:  node.Name,
+		NodeType:  node.Type,
+		StartedAt: startTime.Unix(),
+	}
+
+	var nodeOutput *models.NodeOutput
+
+	// 跳过禁用的节点
+	if node.Disabled {
+		nodeResult.Status = "skipped"
+		nodeResult.Message = "node is disabled"
+		nodeResult.FinishedAt = time.Now().Unix()
+		nodeResult.DurationMs = time.Since(startTime).Milliseconds()
+		return nodeResult, nil
+	}
+
+	// 获取处理器
+	processor, err := models.GetProcessorByType(node.Type, node.Config)
+	if err != nil {
+		nodeResult.Status = "failed"
+		nodeResult.Error = fmt.Sprintf("failed to get processor: %v", err)
+		nodeResult.FinishedAt = time.Now().Unix()
+		nodeResult.DurationMs = time.Since(startTime).Milliseconds()
+		return nodeResult, nil
+	}
+
+	// 执行处理器（带重试）
+	var retries int
+	maxRetries := node.MaxRetries
+	if !node.RetryOnFail {
+		maxRetries = 0
+	}
+
+	for retries <= maxRetries {
+		// 检查是否为分支处理器
+		if branchProcessor, ok := processor.(models.BranchProcessor); ok {
+			output, err := branchProcessor.ProcessWithBranch(e.ctx, wfCtx)
+			if err != nil {
+				if retries < maxRetries {
+					retries++
+					time.Sleep(time.Duration(node.RetryInterval) * time.Second)
+					continue
+				}
+				nodeResult.Status = "failed"
+				nodeResult.Error = err.Error()
+			} else {
+				nodeResult.Status = "success"
+				if output != nil {
+					nodeOutput = output
+					if output.WfCtx != nil {
+						wfCtx = output.WfCtx
+					}
+					nodeResult.Message = output.Message
+					nodeResult.BranchIndex = output.BranchIndex
+					if output.Terminate {
+						nodeResult.Status = "terminated"
+					}
+				}
+			}
+			break
+		}
+
+		// 普通处理器
+		newWfCtx, msg, err := processor.Process(e.ctx, wfCtx)
+		if err != nil {
+			if retries < maxRetries {
+				retries++
+				time.Sleep(time.Duration(node.RetryInterval) * time.Second)
+				continue
+			}
+			nodeResult.Status = "failed"
+			nodeResult.Error = err.Error()
+		} else {
+			nodeResult.Status = "success"
+			nodeResult.Message = msg
+			if newWfCtx != nil {
+				wfCtx = newWfCtx
+
+				// 检测流式输出标记
+				if newWfCtx.Stream && newWfCtx.StreamChan != nil {
+					nodeOutput = &models.NodeOutput{
+						WfCtx:      newWfCtx,
+						Message:    msg,
+						Stream:     true,
+						StreamChan: newWfCtx.StreamChan,
+					}
+				}
+			}
+
+			// 如果事件被 drop（返回 nil 或 Event 为 nil），标记为终止
+			if newWfCtx == nil || newWfCtx.Event == nil {
+				nodeResult.Status = "terminated"
+				nodeResult.Message = msg
+			}
+		}
+		break
+	}
+
+	nodeResult.FinishedAt = time.Now().Unix()
+	nodeResult.DurationMs = time.Since(startTime).Milliseconds()
+
+	logger.Infof("workflow: executed node %s (type=%s) status=%s msg=%s duration=%dms",
+		node.Name, node.Type, nodeResult.Status, nodeResult.Message, nodeResult.DurationMs)
+
+	return nodeResult, nodeOutput
+}
+
+// shouldFollowBranch 判断是否应该走某个分支
+func (e *WorkflowEngine) shouldFollowBranch(nodeID string, outputIndex int, branchResults map[string]*int) bool {
+	branchIndex, hasBranch := branchResults[nodeID]
+	if !hasBranch {
+		// 没有分支结果，说明不是分支节点，只走第一个输出
+		return outputIndex == 0
+	}
+
+	if branchIndex == nil {
+		// branchIndex 为 nil，走默认分支（通常是最后一个）
+		return true
+	}
+
+	// 只走选中的分支
+	return outputIndex == *branchIndex
+}
+
+func (e *WorkflowEngine) saveExecutionRecord(pipeline *models.EventPipeline, wfCtx *models.WorkflowContext, result *models.WorkflowResult, triggerCtx *models.WorkflowTriggerContext, startTime int64, duration int64) {
+	executionID := triggerCtx.RequestID
+	if executionID == "" {
+		executionID = uuid.New().String()
+	}
+
+	execution := &models.EventPipelineExecution{
+		ID:           executionID,
+		PipelineID:   pipeline.ID,
+		PipelineName: pipeline.Name,
+		Mode:         triggerCtx.Mode,
+		Status:       result.Status,
+		ErrorMessage: result.Message,
+		ErrorNode:    result.ErrorNode,
+		CreatedAt:    startTime,
+		FinishedAt:   time.Now().Unix(),
+		DurationMs:   duration,
+		TriggerBy:    triggerCtx.TriggerBy,
+	}
+
+	if wfCtx.Event != nil {
+		execution.EventID = wfCtx.Event.Id
+	}
+
+	if err := execution.SetNodeResults(result.NodeResults); err != nil {
+		logger.Errorf("workflow: failed to set node results: pipeline_id=%d, error=%v", pipeline.ID, err)
+	}
+
+	secretKeys := pipeline.GetSecretKeys()
+	sanitizedEnv := wfCtx.SanitizedEnv(secretKeys)
+	if err := execution.SetEnvSnapshot(sanitizedEnv); err != nil {
+		logger.Errorf("workflow: failed to set env snapshot: pipeline_id=%d, error=%v", pipeline.ID, err)
+	}
+
+	if err := models.CreateEventPipelineExecution(e.ctx, execution); err != nil {
+		logger.Errorf("workflow: failed to save execution record: pipeline_id=%d, error=%v", pipeline.ID, err)
+	}
+}

alert/pipeline/pipeline.go

@@ -5,6 +5,7 @@ import (
 	_ "github.com/ccfos/nightingale/v6/alert/pipeline/processor/callback"
 	_ "github.com/ccfos/nightingale/v6/alert/pipeline/processor/eventdrop"
 	_ "github.com/ccfos/nightingale/v6/alert/pipeline/processor/eventupdate"
+	_ "github.com/ccfos/nightingale/v6/alert/pipeline/processor/logic"
 	_ "github.com/ccfos/nightingale/v6/alert/pipeline/processor/relabel"
 )
 

alert/pipeline/processor/aisummary/ai_summary.go

@@ -55,23 +55,24 @@ func (c *AISummaryConfig) Init(settings interface{}) (models.Processor, error) {
 	return result, err
 }
 
-func (c *AISummaryConfig) Process(ctx *ctx.Context, event *models.AlertCurEvent) (*models.AlertCurEvent, string, error) {
+func (c *AISummaryConfig) Process(ctx *ctx.Context, wfCtx *models.WorkflowContext) (*models.WorkflowContext, string, error) {
+	event := wfCtx.Event
 	if c.Client == nil {
 		if err := c.initHTTPClient(); err != nil {
-			return event, "", fmt.Errorf("failed to initialize HTTP client: %v processor: %v", err, c)
+			return wfCtx, "", fmt.Errorf("failed to initialize HTTP client: %v processor: %v", err, c)
 		}
 	}
 
 	// 准备告警事件信息
-	eventInfo, err := c.prepareEventInfo(event)
+	eventInfo, err := c.prepareEventInfo(wfCtx)
 	if err != nil {
-		return event, "", fmt.Errorf("failed to prepare event info: %v processor: %v", err, c)
+		return wfCtx, "", fmt.Errorf("failed to prepare event info: %v processor: %v", err, c)
 	}
 
 	// 调用AI模型生成总结
 	summary, err := c.generateAISummary(eventInfo)
 	if err != nil {
-		return event, "", fmt.Errorf("failed to generate AI summary: %v processor: %v", err, c)
+		return wfCtx, "", fmt.Errorf("failed to generate AI summary: %v processor: %v", err, c)
 	}
 
 	// 将总结添加到annotations字段
@@ -83,11 +84,11 @@ func (c *AISummaryConfig) Process(ctx *ctx.Context, event *models.AlertCurEvent)
 	// 更新Annotations字段
 	b, err := json.Marshal(event.AnnotationsJSON)
 	if err != nil {
-		return event, "", fmt.Errorf("failed to marshal annotations: %v processor: %v", err, c)
+		return wfCtx, "", fmt.Errorf("failed to marshal annotations: %v processor: %v", err, c)
 	}
 	event.Annotations = string(b)
 
-	return event, "", nil
+	return wfCtx, "", nil
 }
 
 func (c *AISummaryConfig) initHTTPClient() error {
@@ -110,9 +111,10 @@ func (c *AISummaryConfig) initHTTPClient() error {
 	return nil
 }
 
-func (c *AISummaryConfig) prepareEventInfo(event *models.AlertCurEvent) (string, error) {
+func (c *AISummaryConfig) prepareEventInfo(wfCtx *models.WorkflowContext) (string, error) {
 	var defs = []string{
-		"{{$event := .}}",
+		"{{$event := .Event}}",
+		"{{$env := .Env}}",
 	}
 
 	text := strings.Join(append(defs, c.PromptTemplate), "")
@@ -122,7 +124,7 @@ func (c *AISummaryConfig) prepareEventInfo(event *models.AlertCurEvent) (string,
 	}
 
 	var body bytes.Buffer
-	err = t.Execute(&body, event)
+	err = t.Execute(&body, wfCtx)
 	if err != nil {
 		return "", fmt.Errorf("failed to execute prompt template: %v", err)
 	}

alert/pipeline/processor/aisummary/ai_summary_test.go

@@ -42,8 +42,14 @@ func TestAISummaryConfig_Process(t *testing.T) {
 		},
 	}
 
+	// 创建 WorkflowContext
+	wfCtx := &models.WorkflowContext{
+		Event: event,
+		Env:   map[string]string{},
+	}
+
 	// 测试模板处理
-	eventInfo, err := config.prepareEventInfo(event)
+	eventInfo, err := config.prepareEventInfo(wfCtx)
 	assert.NoError(t, err)
 	assert.Contains(t, eventInfo, "Test Rule")
 	assert.Contains(t, eventInfo, "1")
@@ -54,18 +60,18 @@ func TestAISummaryConfig_Process(t *testing.T) {
 	assert.NotNil(t, processor)
 
 	// 测试处理函数
-	result, _, err := processor.Process(&ctx.Context{}, event)
+	result, _, err := processor.Process(&ctx.Context{}, wfCtx)
 	assert.NoError(t, err)
 	assert.NotNil(t, result)
-	assert.NotEmpty(t, result.AnnotationsJSON["ai_summary"])
+	assert.NotEmpty(t, result.Event.AnnotationsJSON["ai_summary"])
 
 	// 展示处理结果
 	t.Log("\n=== 处理结果 ===")
-	t.Logf("告警规则: %s", result.RuleName)
-	t.Logf("严重程度: %d", result.Severity)
-	t.Logf("标签: %v", result.TagsMap)
-	t.Logf("原始注释: %v", result.AnnotationsJSON["description"])
-	t.Logf("AI总结: %s", result.AnnotationsJSON["ai_summary"])
+	t.Logf("告警规则: %s", result.Event.RuleName)
+	t.Logf("严重程度: %d", result.Event.Severity)
+	t.Logf("标签: %v", result.Event.TagsMap)
+	t.Logf("原始注释: %v", result.Event.AnnotationsJSON["description"])
+	t.Logf("AI总结: %s", result.Event.AnnotationsJSON["ai_summary"])
 }
 
 func TestConvertCustomParam(t *testing.T) {

alert/pipeline/processor/callback/callback.go

@@ -43,7 +43,8 @@ func (c *CallbackConfig) Init(settings interface{}) (models.Processor, error) {
 	return result, err
 }
 
-func (c *CallbackConfig) Process(ctx *ctx.Context, event *models.AlertCurEvent) (*models.AlertCurEvent, string, error) {
+func (c *CallbackConfig) Process(ctx *ctx.Context, wfCtx *models.WorkflowContext) (*models.WorkflowContext, string, error) {
+	event := wfCtx.Event
 	if c.Client == nil {
 		transport := &http.Transport{
 			TLSClientConfig: &tls.Config{InsecureSkipVerify: c.SkipSSLVerify},
@@ -52,7 +53,7 @@ func (c *CallbackConfig) Process(ctx *ctx.Context, event *models.AlertCurEvent)
 		if c.Proxy != "" {
 			proxyURL, err := url.Parse(c.Proxy)
 			if err != nil {
-				return event, "", fmt.Errorf("failed to parse proxy url: %v processor: %v", err, c)
+				return wfCtx, "", fmt.Errorf("failed to parse proxy url: %v processor: %v", err, c)
 			} else {
 				transport.Proxy = http.ProxyURL(proxyURL)
 			}
@@ -72,12 +73,12 @@ func (c *CallbackConfig) Process(ctx *ctx.Context, event *models.AlertCurEvent)
 
 	body, err := json.Marshal(event)
 	if err != nil {
-		return event, "", fmt.Errorf("failed to marshal event: %v processor: %v", err, c)
+		return wfCtx, "", fmt.Errorf("failed to marshal event: %v processor: %v", err, c)
 	}
 
 	req, err := http.NewRequest("POST", c.URL, strings.NewReader(string(body)))
 	if err != nil {
-		return event, "", fmt.Errorf("failed to create request: %v processor: %v", err, c)
+		return wfCtx, "", fmt.Errorf("failed to create request: %v processor: %v", err, c)
 	}
 
 	for k, v := range headers {
@@ -90,14 +91,14 @@ func (c *CallbackConfig) Process(ctx *ctx.Context, event *models.AlertCurEvent)
 
 	resp, err := c.Client.Do(req)
 	if err != nil {
-		return event, "", fmt.Errorf("failed to send request: %v processor: %v", err, c)
+		return wfCtx, "", fmt.Errorf("failed to send request: %v processor: %v", err, c)
 	}
 
 	b, err := io.ReadAll(resp.Body)
 	if err != nil {
-		return event, "", fmt.Errorf("failed to read response body: %v processor: %v", err, c)
+		return wfCtx, "", fmt.Errorf("failed to read response body: %v processor: %v", err, c)
 	}
 
 	logger.Debugf("callback processor response body: %s", string(b))
-	return event, "callback success", nil
+	return wfCtx, "callback success", nil
 }

alert/pipeline/processor/eventdrop/event_drop.go

@@ -26,27 +26,29 @@ func (c *EventDropConfig) Init(settings interface{}) (models.Processor, error) {
 	return result, err
 }
 
-func (c *EventDropConfig) Process(ctx *ctx.Context, event *models.AlertCurEvent) (*models.AlertCurEvent, string, error) {
+func (c *EventDropConfig) Process(ctx *ctx.Context, wfCtx *models.WorkflowContext) (*models.WorkflowContext, string, error) {
 	// 使用背景是可以根据此处理器，实现对事件进行更加灵活的过滤的逻辑
 	// 在标签过滤和属性过滤都不满足需求时可以使用
 	// 如果模板执行结果为 true，则删除该事件
+	event := wfCtx.Event
 
 	var defs = []string{
-		"{{ $event := . }}",
-		"{{ $labels := .TagsMap }}",
-		"{{ $value := .TriggerValue }}",
+		"{{ $event := .Event }}",
+		"{{ $labels := .Event.TagsMap }}",
+		"{{ $value := .Event.TriggerValue }}",
+		"{{ $env := .Env }}",
 	}
 
 	text := strings.Join(append(defs, c.Content), "")
 
 	tpl, err := texttemplate.New("eventdrop").Funcs(tplx.TemplateFuncMap).Parse(text)
 	if err != nil {
-		return event, "", fmt.Errorf("processor failed to parse template: %v processor: %v", err, c)
+		return wfCtx, "", fmt.Errorf("processor failed to parse template: %v processor: %v", err, c)
 	}
 
 	var body bytes.Buffer
-	if err = tpl.Execute(&body, event); err != nil {
-		return event, "", fmt.Errorf("processor failed to execute template: %v processor: %v", err, c)
+	if err = tpl.Execute(&body, wfCtx); err != nil {
+		return wfCtx, "", fmt.Errorf("processor failed to execute template: %v processor: %v", err, c)
 	}
 
 	result := strings.TrimSpace(body.String())
@@ -56,5 +58,5 @@ func (c *EventDropConfig) Process(ctx *ctx.Context, event *models.AlertCurEvent)
 		return nil, "drop event success", nil
 	}
 
-	return event, "drop event failed", nil
+	return wfCtx, "drop event failed", nil
 }

alert/pipeline/processor/eventupdate/event_update.go

@@ -31,7 +31,8 @@ func (c *EventUpdateConfig) Init(settings interface{}) (models.Processor, error)
 	return result, err
 }
 
-func (c *EventUpdateConfig) Process(ctx *ctx.Context, event *models.AlertCurEvent) (*models.AlertCurEvent, string, error) {
+func (c *EventUpdateConfig) Process(ctx *ctx.Context, wfCtx *models.WorkflowContext) (*models.WorkflowContext, string, error) {
+	event := wfCtx.Event
 	if c.Client == nil {
 		transport := &http.Transport{
 			TLSClientConfig: &tls.Config{InsecureSkipVerify: c.SkipSSLVerify},
@@ -40,7 +41,7 @@ func (c *EventUpdateConfig) Process(ctx *ctx.Context, event *models.AlertCurEven
 		if c.Proxy != "" {
 			proxyURL, err := url.Parse(c.Proxy)
 			if err != nil {
-				return event, "", fmt.Errorf("failed to parse proxy url: %v processor: %v", err, c)
+				return wfCtx, "", fmt.Errorf("failed to parse proxy url: %v processor: %v", err, c)
 			} else {
 				transport.Proxy = http.ProxyURL(proxyURL)
 			}
@@ -60,12 +61,12 @@ func (c *EventUpdateConfig) Process(ctx *ctx.Context, event *models.AlertCurEven
 
 	body, err := json.Marshal(event)
 	if err != nil {
-		return event, "", fmt.Errorf("failed to marshal event: %v processor: %v", err, c)
+		return wfCtx, "", fmt.Errorf("failed to marshal event: %v processor: %v", err, c)
 	}
 
 	req, err := http.NewRequest("POST", c.URL, strings.NewReader(string(body)))
 	if err != nil {
-		return event, "", fmt.Errorf("failed to create request: %v processor: %v", err, c)
+		return wfCtx, "", fmt.Errorf("failed to create request: %v processor: %v", err, c)
 	}
 
 	for k, v := range headers {
@@ -78,7 +79,7 @@ func (c *EventUpdateConfig) Process(ctx *ctx.Context, event *models.AlertCurEven
 
 	resp, err := c.Client.Do(req)
 	if err != nil {
-		return event, "", fmt.Errorf("failed to send request: %v processor: %v", err, c)
+		return wfCtx, "", fmt.Errorf("failed to send request: %v processor: %v", err, c)
 	}
 
 	b, err := io.ReadAll(resp.Body)
@@ -89,8 +90,8 @@ func (c *EventUpdateConfig) Process(ctx *ctx.Context, event *models.AlertCurEven
 
 	err = json.Unmarshal(b, &event)
 	if err != nil {
-		return event, "", fmt.Errorf("failed to unmarshal response body: %v processor: %v", err, c)
+		return wfCtx, "", fmt.Errorf("failed to unmarshal response body: %v processor: %v", err, c)
 	}
 
-	return event, "", nil
+	return wfCtx, "", nil
 }

alert/pipeline/processor/logic/if.go

@@ -0,0 +1,197 @@
+package logic
+
+import (
+	"bytes"
+	"fmt"
+	"strings"
+	"text/template"
+
+	alertCommon "github.com/ccfos/nightingale/v6/alert/common"
+	"github.com/ccfos/nightingale/v6/alert/pipeline/processor/common"
+	"github.com/ccfos/nightingale/v6/models"
+	"github.com/ccfos/nightingale/v6/pkg/ctx"
+	"github.com/ccfos/nightingale/v6/pkg/tplx"
+)
+
+// 判断模式常量
+const (
+	ConditionModeExpression = "expression" // 表达式模式（默认）
+	ConditionModeTags       = "tags"       // 标签/属性模式
+)
+
+// IfConfig If 条件处理器配置
+type IfConfig struct {
+	// 判断模式：expression（表达式）或 tags（标签/属性）
+	Mode string `json:"mode,omitempty"`
+
+	// 表达式模式配置
+	// 条件表达式（支持 Go 模板语法）
+	// 例如：{{ if eq .Severity 1 }}true{{ end }}
+	Condition string `json:"condition,omitempty"`
+
+	// 标签/属性模式配置
+	LabelKeys  []models.TagFilter `json:"label_keys,omitempty"` // 适用标签
+	Attributes []models.TagFilter `json:"attributes,omitempty"` // 适用属性
+
+	// 内部使用，解析后的过滤器
+	parsedLabelKeys  []models.TagFilter `json:"-"`
+	parsedAttributes []models.TagFilter `json:"-"`
+}
+
+func init() {
+	models.RegisterProcessor("logic.if", &IfConfig{})
+}
+
+func (c *IfConfig) Init(settings interface{}) (models.Processor, error) {
+	result, err := common.InitProcessor[*IfConfig](settings)
+	if err != nil {
+		return nil, err
+	}
+
+	// 解析标签过滤器
+	if len(result.LabelKeys) > 0 {
+		// Deep copy to avoid concurrent map writes on cached objects
+		labelKeysCopy := make([]models.TagFilter, len(result.LabelKeys))
+		copy(labelKeysCopy, result.LabelKeys)
+		for i := range labelKeysCopy {
+			if labelKeysCopy[i].Func == "" {
+				labelKeysCopy[i].Func = labelKeysCopy[i].Op
+			}
+		}
+		result.parsedLabelKeys, err = models.ParseTagFilter(labelKeysCopy)
+		if err != nil {
+			return nil, fmt.Errorf("failed to parse label_keys: %v", err)
+		}
+	}
+
+	// 解析属性过滤器
+	if len(result.Attributes) > 0 {
+		// Deep copy to avoid concurrent map writes on cached objects
+		attributesCopy := make([]models.TagFilter, len(result.Attributes))
+		copy(attributesCopy, result.Attributes)
+		for i := range attributesCopy {
+			if attributesCopy[i].Func == "" {
+				attributesCopy[i].Func = attributesCopy[i].Op
+			}
+		}
+		result.parsedAttributes, err = models.ParseTagFilter(attributesCopy)
+		if err != nil {
+			return nil, fmt.Errorf("failed to parse attributes: %v", err)
+		}
+	}
+
+	return result, nil
+}
+
+// Process 实现 Processor 接口（兼容旧模式）
+func (c *IfConfig) Process(ctx *ctx.Context, wfCtx *models.WorkflowContext) (*models.WorkflowContext, string, error) {
+	result, err := c.evaluateCondition(wfCtx)
+	if err != nil {
+		return wfCtx, "", fmt.Errorf("if processor: failed to evaluate condition: %v", err)
+	}
+
+	if result {
+		return wfCtx, "condition matched (true branch)", nil
+	}
+	return wfCtx, "condition not matched (false branch)", nil
+}
+
+// ProcessWithBranch 实现 BranchProcessor 接口
+func (c *IfConfig) ProcessWithBranch(ctx *ctx.Context, wfCtx *models.WorkflowContext) (*models.NodeOutput, error) {
+	result, err := c.evaluateCondition(wfCtx)
+	if err != nil {
+		return nil, fmt.Errorf("if processor: failed to evaluate condition: %v", err)
+	}
+
+	output := &models.NodeOutput{
+		WfCtx: wfCtx,
+	}
+
+	if result {
+		// 条件为 true，走输出 0（true 分支）
+		branchIndex := 0
+		output.BranchIndex = &branchIndex
+		output.Message = "condition matched (true branch)"
+	} else {
+		// 条件为 false，走输出 1（false 分支）
+		branchIndex := 1
+		output.BranchIndex = &branchIndex
+		output.Message = "condition not matched (false branch)"
+	}
+
+	return output, nil
+}
+
+// evaluateCondition 评估条件
+func (c *IfConfig) evaluateCondition(wfCtx *models.WorkflowContext) (bool, error) {
+	mode := c.Mode
+	if mode == "" {
+		mode = ConditionModeExpression // 默认表达式模式
+	}
+
+	switch mode {
+	case ConditionModeTags:
+		return c.evaluateTagsCondition(wfCtx.Event)
+	default:
+		return c.evaluateExpressionCondition(wfCtx)
+	}
+}
+
+// evaluateExpressionCondition 评估表达式条件
+func (c *IfConfig) evaluateExpressionCondition(wfCtx *models.WorkflowContext) (bool, error) {
+	if c.Condition == "" {
+		return true, nil
+	}
+
+	// 构建模板数据
+	var defs = []string{
+		"{{ $event := .Event }}",
+		"{{ $labels := .Event.TagsMap }}",
+		"{{ $value := .Event.TriggerValue }}",
+		"{{ $env := .Env }}",
+	}
+
+	text := strings.Join(append(defs, c.Condition), "")
+
+	tpl, err := template.New("if_condition").Funcs(tplx.TemplateFuncMap).Parse(text)
+	if err != nil {
+		return false, err
+	}
+
+	var buf bytes.Buffer
+	if err = tpl.Execute(&buf, wfCtx); err != nil {
+		return false, err
+	}
+
+	result := strings.TrimSpace(strings.ToLower(buf.String()))
+	return result == "true" || result == "1", nil
+}
+
+// evaluateTagsCondition 评估标签/属性条件
+func (c *IfConfig) evaluateTagsCondition(event *models.AlertCurEvent) (bool, error) {
+	// 如果没有配置任何过滤条件，默认返回 true
+	if len(c.parsedLabelKeys) == 0 && len(c.parsedAttributes) == 0 {
+		return true, nil
+	}
+
+	// 匹配标签 (TagsMap)
+	if len(c.parsedLabelKeys) > 0 {
+		tagsMap := event.TagsMap
+		if tagsMap == nil {
+			tagsMap = make(map[string]string)
+		}
+		if !alertCommon.MatchTags(tagsMap, c.parsedLabelKeys) {
+			return false, nil
+		}
+	}
+
+	// 匹配属性 (JsonTagsAndValue - 所有 JSON 字段)
+	if len(c.parsedAttributes) > 0 {
+		attributesMap := event.JsonTagsAndValue()
+		if !alertCommon.MatchTags(attributesMap, c.parsedAttributes) {
+			return false, nil
+		}
+	}
+
+	return true, nil
+}

alert/pipeline/processor/logic/switch.go

@@ -0,0 +1,224 @@
+package logic
+
+import (
+	"bytes"
+	"fmt"
+	"strings"
+	"text/template"
+
+	alertCommon "github.com/ccfos/nightingale/v6/alert/common"
+	"github.com/ccfos/nightingale/v6/alert/pipeline/processor/common"
+	"github.com/ccfos/nightingale/v6/models"
+	"github.com/ccfos/nightingale/v6/pkg/ctx"
+	"github.com/ccfos/nightingale/v6/pkg/tplx"
+)
+
+// SwitchCase Switch 分支定义
+type SwitchCase struct {
+	// 判断模式：expression（表达式）或 tags（标签/属性）
+	Mode string `json:"mode,omitempty"`
+
+	// 表达式模式配置
+	// 条件表达式（支持 Go 模板语法）
+	Condition string `json:"condition,omitempty"`
+
+	// 标签/属性模式配置
+	LabelKeys  []models.TagFilter `json:"label_keys,omitempty"` // 适用标签
+	Attributes []models.TagFilter `json:"attributes,omitempty"` // 适用属性
+
+	// 分支名称（可选，用于日志）
+	Name string `json:"name,omitempty"`
+
+	// 内部使用，解析后的过滤器
+	parsedLabelKeys  []models.TagFilter `json:"-"`
+	parsedAttributes []models.TagFilter `json:"-"`
+}
+
+// SwitchConfig Switch 多分支处理器配置
+type SwitchConfig struct {
+	// 分支条件列表
+	// 按顺序匹配，第一个为 true 的分支将被选中
+	Cases []SwitchCase `json:"cases"`
+	// 是否允许多个分支同时匹配（默认 false，只走第一个匹配的）
+	AllowMultiple bool `json:"allow_multiple,omitempty"`
+}
+
+func init() {
+	models.RegisterProcessor("logic.switch", &SwitchConfig{})
+}
+
+func (c *SwitchConfig) Init(settings interface{}) (models.Processor, error) {
+	result, err := common.InitProcessor[*SwitchConfig](settings)
+	if err != nil {
+		return nil, err
+	}
+
+	// 解析每个 case 的标签和属性过滤器
+	for i := range result.Cases {
+		if len(result.Cases[i].LabelKeys) > 0 {
+			// Deep copy to avoid concurrent map writes on cached objects
+			labelKeysCopy := make([]models.TagFilter, len(result.Cases[i].LabelKeys))
+			copy(labelKeysCopy, result.Cases[i].LabelKeys)
+			for j := range labelKeysCopy {
+				if labelKeysCopy[j].Func == "" {
+					labelKeysCopy[j].Func = labelKeysCopy[j].Op
+				}
+			}
+			result.Cases[i].parsedLabelKeys, err = models.ParseTagFilter(labelKeysCopy)
+			if err != nil {
+				return nil, fmt.Errorf("failed to parse label_keys for case[%d]: %v", i, err)
+			}
+		}
+
+		if len(result.Cases[i].Attributes) > 0 {
+			// Deep copy to avoid concurrent map writes on cached objects
+			attributesCopy := make([]models.TagFilter, len(result.Cases[i].Attributes))
+			copy(attributesCopy, result.Cases[i].Attributes)
+			for j := range attributesCopy {
+				if attributesCopy[j].Func == "" {
+					attributesCopy[j].Func = attributesCopy[j].Op
+				}
+			}
+			result.Cases[i].parsedAttributes, err = models.ParseTagFilter(attributesCopy)
+			if err != nil {
+				return nil, fmt.Errorf("failed to parse attributes for case[%d]: %v", i, err)
+			}
+		}
+	}
+
+	return result, nil
+}
+
+// Process 实现 Processor 接口（兼容旧模式）
+func (c *SwitchConfig) Process(ctx *ctx.Context, wfCtx *models.WorkflowContext) (*models.WorkflowContext, string, error) {
+	index, caseName, err := c.evaluateCases(wfCtx)
+	if err != nil {
+		return wfCtx, "", fmt.Errorf("switch processor: failed to evaluate cases: %v", err)
+	}
+
+	if index >= 0 {
+		if caseName != "" {
+			return wfCtx, fmt.Sprintf("matched case[%d]: %s", index, caseName), nil
+		}
+		return wfCtx, fmt.Sprintf("matched case[%d]", index), nil
+	}
+
+	// 走默认分支（最后一个输出）
+	return wfCtx, "no case matched, using default branch", nil
+}
+
+// ProcessWithBranch 实现 BranchProcessor 接口
+func (c *SwitchConfig) ProcessWithBranch(ctx *ctx.Context, wfCtx *models.WorkflowContext) (*models.NodeOutput, error) {
+	index, caseName, err := c.evaluateCases(wfCtx)
+	if err != nil {
+		return nil, fmt.Errorf("switch processor: failed to evaluate cases: %v", err)
+	}
+
+	output := &models.NodeOutput{
+		WfCtx: wfCtx,
+	}
+
+	if index >= 0 {
+		output.BranchIndex = &index
+		if caseName != "" {
+			output.Message = fmt.Sprintf("matched case[%d]: %s", index, caseName)
+		} else {
+			output.Message = fmt.Sprintf("matched case[%d]", index)
+		}
+	} else {
+		// 默认分支的索引是 cases 数量（即最后一个输出端口）
+		defaultIndex := len(c.Cases)
+		output.BranchIndex = &defaultIndex
+		output.Message = "no case matched, using default branch"
+	}
+
+	return output, nil
+}
+
+// evaluateCases 评估所有分支条件
+// 返回匹配的分支索引和分支名称，如果没有匹配返回 -1
+func (c *SwitchConfig) evaluateCases(wfCtx *models.WorkflowContext) (int, string, error) {
+	for i := range c.Cases {
+		matched, err := c.evaluateCaseCondition(&c.Cases[i], wfCtx)
+		if err != nil {
+			return -1, "", fmt.Errorf("case[%d] evaluation error: %v", i, err)
+		}
+		if matched {
+			return i, c.Cases[i].Name, nil
+		}
+	}
+	return -1, "", nil
+}
+
+// evaluateCaseCondition 评估单个分支条件
+func (c *SwitchConfig) evaluateCaseCondition(caseItem *SwitchCase, wfCtx *models.WorkflowContext) (bool, error) {
+	mode := caseItem.Mode
+	if mode == "" {
+		mode = ConditionModeExpression // 默认表达式模式
+	}
+
+	switch mode {
+	case ConditionModeTags:
+		return c.evaluateTagsCondition(caseItem, wfCtx.Event)
+	default:
+		return c.evaluateExpressionCondition(caseItem.Condition, wfCtx)
+	}
+}
+
+// evaluateExpressionCondition 评估表达式条件
+func (c *SwitchConfig) evaluateExpressionCondition(condition string, wfCtx *models.WorkflowContext) (bool, error) {
+	if condition == "" {
+		return false, nil
+	}
+
+	var defs = []string{
+		"{{ $event := .Event }}",
+		"{{ $labels := .Event.TagsMap }}",
+		"{{ $value := .Event.TriggerValue }}",
+		"{{ $env := .Env }}",
+	}
+
+	text := strings.Join(append(defs, condition), "")
+
+	tpl, err := template.New("switch_condition").Funcs(tplx.TemplateFuncMap).Parse(text)
+	if err != nil {
+		return false, err
+	}
+
+	var buf bytes.Buffer
+	if err = tpl.Execute(&buf, wfCtx); err != nil {
+		return false, err
+	}
+
+	result := strings.TrimSpace(strings.ToLower(buf.String()))
+	return result == "true" || result == "1", nil
+}
+
+// evaluateTagsCondition 评估标签/属性条件
+func (c *SwitchConfig) evaluateTagsCondition(caseItem *SwitchCase, event *models.AlertCurEvent) (bool, error) {
+	// 如果没有配置任何过滤条件，默认返回 false（不匹配）
+	if len(caseItem.parsedLabelKeys) == 0 && len(caseItem.parsedAttributes) == 0 {
+		return false, nil
+	}
+
+	// 匹配标签 (TagsMap)
+	if len(caseItem.parsedLabelKeys) > 0 {
+		tagsMap := event.TagsMap
+		if tagsMap == nil {
+			tagsMap = make(map[string]string)
+		}
+		if !alertCommon.MatchTags(tagsMap, caseItem.parsedLabelKeys) {
+			return false, nil
+		}
+	}
+
+	// 匹配属性 (JsonTagsAndValue - 所有 JSON 字段)
+	if len(caseItem.parsedAttributes) > 0 {
+		attributesMap := event.JsonTagsAndValue()
+		if !alertCommon.MatchTags(attributesMap, caseItem.parsedAttributes) {
+			return false, nil
+		}
+	}
+
+	return true, nil
+}

alert/pipeline/processor/relabel/relabel.go

@@ -42,7 +42,7 @@ func (r *RelabelConfig) Init(settings interface{}) (models.Processor, error) {
 	return result, err
 }
 
-func (r *RelabelConfig) Process(ctx *ctx.Context, event *models.AlertCurEvent) (*models.AlertCurEvent, string, error) {
+func (r *RelabelConfig) Process(ctx *ctx.Context, wfCtx *models.WorkflowContext) (*models.WorkflowContext, string, error) {
 	sourceLabels := make([]model.LabelName, len(r.SourceLabels))
 	for i := range r.SourceLabels {
 		sourceLabels[i] = model.LabelName(strings.ReplaceAll(r.SourceLabels[i], ".", REPLACE_DOT))
@@ -63,8 +63,8 @@ func (r *RelabelConfig) Process(ctx *ctx.Context, event *models.AlertCurEvent) (
 		},
 	}
 
-	EventRelabel(event, relabelConfigs)
-	return event, "", nil
+	EventRelabel(wfCtx.Event, relabelConfigs)
+	return wfCtx, "", nil
 }
 
 func EventRelabel(event *models.AlertCurEvent, relabelConfigs []*pconf.RelabelConfig) {

alert/sender/callback.go

@@ -135,9 +135,7 @@ func (c *DefaultCallBacker) CallBack(ctx CallBackContext) {
 
 func doSendAndRecord(ctx *ctx.Context, url, token string, body interface{}, channel string,
 	stats *astats.Stats, events []*models.AlertCurEvent) {
-	start := time.Now()
 	res, err := doSend(url, body, channel, stats)
-	res = fmt.Sprintf("duration: %d ms %s", time.Since(start).Milliseconds(), res)
 	NotifyRecord(ctx, events, 0, channel, token, res, err)
 }
 
@@ -171,11 +169,11 @@ func doSend(url string, body interface{}, channel string, stats *astats.Stats) (
 
 	start := time.Now()
 	res, code, err := poster.PostJSON(url, time.Second*5, body, 3)
-	res = []byte(fmt.Sprintf("duration: %d ms %s", time.Since(start).Milliseconds(), res))
+	res = []byte(fmt.Sprintf("duration: %d ms status_code:%d, response:%s", time.Since(start).Milliseconds(), code, string(res)))
 	if err != nil {
 		logger.Errorf("%s_sender: result=fail url=%s code=%d error=%v req:%v response=%s", channel, url, code, err, body, string(res))
 		stats.AlertNotifyErrorTotal.WithLabelValues(channel).Inc()
-		return "", err
+		return string(res), err
 	}
 
 	logger.Infof("%s_sender: result=succ url=%s code=%d req:%v response=%s", channel, url, code, body, string(res))

alert/sender/plugin.go

@@ -89,7 +89,7 @@ func alertingCallScript(ctx *ctx.Context, stdinBytes []byte, notifyScript models
 	err, isTimeout := sys.WrapTimeout(cmd, time.Duration(config.Timeout)*time.Second)
 
 	res := buf.String()
-	res = fmt.Sprintf("duration: %d ms %s", time.Since(start).Milliseconds(), res)
+	res = fmt.Sprintf("send_time: %s duration: %d ms %s", time.Now().Format("2006-01-02 15:04:05"), time.Since(start).Milliseconds(), res)
 
 	// 截断超出长度的输出
 	if len(res) > 512 {

alert/sender/webhook.go

@@ -119,11 +119,11 @@ func sendWebhook(webhook *models.Webhook, event interface{}, stats *astats.Stats
 
 	if resp.StatusCode == 429 {
 		logger.Errorf("event_%s_fail, url: %s, response code: %d, body: %s event:%s", channel, conf.Url, resp.StatusCode, string(body), string(bs))
-		return true, string(body), fmt.Errorf("status code is 429")
+		return true, fmt.Sprintf("status_code:%d, response:%s", resp.StatusCode, string(body)), fmt.Errorf("status code is 429")
 	}
 
 	logger.Debugf("event_%s_succ, url: %s, response code: %d, body: %s event:%s", channel, conf.Url, resp.StatusCode, string(body), string(bs))
-	return false, string(body), nil
+	return false, fmt.Sprintf("status_code:%d, response:%s", resp.StatusCode, string(body)), nil
 }
 
 func SingleSendWebhooks(ctx *ctx.Context, webhooks map[string]*models.Webhook, event *models.AlertCurEvent, stats *astats.Stats) {
@@ -132,7 +132,7 @@ func SingleSendWebhooks(ctx *ctx.Context, webhooks map[string]*models.Webhook, e
 		for retryCount < 3 {
 			start := time.Now()
 			needRetry, res, err := sendWebhook(conf, event, stats)
-			res = fmt.Sprintf("duration: %d ms %s", time.Since(start).Milliseconds(), res)
+			res = fmt.Sprintf("send_time: %s duration: %d ms %s", time.Now().Format("2006-01-02 15:04:05"), time.Since(start).Milliseconds(), res)
 			NotifyRecord(ctx, []*models.AlertCurEvent{event}, 0, "webhook", conf.Url, res, err)
 			if !needRetry {
 				break
@@ -204,7 +204,7 @@ func StartConsumer(ctx *ctx.Context, queue *WebhookQueue, popSize int, webhook *
 			for retryCount < webhook.RetryCount {
 				start := time.Now()
 				needRetry, res, err := sendWebhook(webhook, events, stats)
-				res = fmt.Sprintf("duration: %d ms %s", time.Since(start).Milliseconds(), res)
+				res = fmt.Sprintf("send_time: %s duration: %d ms %s", time.Now().Format("2006-01-02 15:04:05"), time.Since(start).Milliseconds(), res)
 				go NotifyRecord(ctx, events, 0, "webhook", webhook.Url, res, err)
 				if !needRetry {
 					break

center/router/router.go

@@ -211,8 +211,8 @@ func (rt *Router) Config(r *gin.Engine) {
 			pages.GET("/datasource/brief", rt.auth(), rt.user(), rt.datasourceBriefs)
 			pages.POST("/datasource/query", rt.auth(), rt.user(), rt.datasourceQuery)
 
-			pages.POST("/ds-query", rt.auth(), rt.QueryData)
-			pages.POST("/logs-query", rt.auth(), rt.QueryLogV2)
+			pages.POST("/ds-query", rt.auth(), rt.user(), rt.QueryData)
+			pages.POST("/logs-query", rt.auth(), rt.user(), rt.QueryLogV2)
 
 			pages.POST("/tdengine-databases", rt.auth(), rt.tdengineDatabases)
 			pages.POST("/tdengine-tables", rt.auth(), rt.tdengineTables)
@@ -558,6 +558,19 @@ func (rt *Router) Config(r *gin.Engine) {
 		pages.POST("/event-pipeline-tryrun", rt.auth(), rt.user(), rt.perm("/event-pipelines"), rt.tryRunEventPipeline)
 		pages.POST("/event-processor-tryrun", rt.auth(), rt.user(), rt.perm("/event-pipelines"), rt.tryRunEventProcessor)
 
+		// API 触发工作流
+		pages.POST("/event-pipeline/:id/trigger", rt.auth(), rt.user(), rt.perm("/event-pipelines"), rt.triggerEventPipelineByAPI)
+		// SSE 流式执行工作流
+		pages.POST("/event-pipeline/:id/stream", rt.auth(), rt.user(), rt.perm("/event-pipelines"), rt.streamEventPipeline)
+
+		// 事件Pipeline执行记录路由
+		pages.GET("/event-pipeline-executions", rt.auth(), rt.user(), rt.perm("/event-pipelines"), rt.listAllEventPipelineExecutions)
+		pages.GET("/event-pipeline/:id/executions", rt.auth(), rt.user(), rt.perm("/event-pipelines"), rt.listEventPipelineExecutions)
+		pages.GET("/event-pipeline/:id/execution/:exec_id", rt.auth(), rt.user(), rt.perm("/event-pipelines"), rt.getEventPipelineExecution)
+		pages.GET("/event-pipeline-execution/:exec_id", rt.auth(), rt.user(), rt.perm("/event-pipelines"), rt.getEventPipelineExecution)
+		pages.GET("/event-pipeline/:id/execution-stats", rt.auth(), rt.user(), rt.perm("/event-pipelines"), rt.getEventPipelineExecutionStats)
+		pages.POST("/event-pipeline-executions/clean", rt.auth(), rt.user(), rt.admin(), rt.cleanEventPipelineExecutions)
+
 		pages.POST("/notify-channel-configs", rt.auth(), rt.user(), rt.perm("/notification-channels/add"), rt.notifyChannelsAdd)
 		pages.DELETE("/notify-channel-configs", rt.auth(), rt.user(), rt.perm("/notification-channels/del"), rt.notifyChannelsDel)
 		pages.PUT("/notify-channel-config/:id", rt.auth(), rt.user(), rt.perm("/notification-channels/put"), rt.notifyChannelPut)
@@ -690,6 +703,8 @@ func (rt *Router) Config(r *gin.Engine) {
 			service.GET("/message-templates", rt.messageTemplateGets)
 
 			service.GET("/event-pipelines", rt.eventPipelinesListByService)
+			service.POST("/event-pipeline/:id/trigger", rt.triggerEventPipelineByService)
+			service.POST("/event-pipeline/:id/stream", rt.streamEventPipelineByService)
 
 			// 手机号加密存储配置接口
 			service.POST("/users/phone/encrypt", rt.usersPhoneEncrypt)

center/router/router_datasource.go

@@ -293,11 +293,15 @@ func DatasourceCheck(c *gin.Context, ds models.Datasource) error {
 		}
 	}
 
+	// 使用 TLS 配置（支持 mTLS）
+	tlsConfig, err := ds.HTTPJson.TLS.TLSConfig()
+	if err != nil {
+		return fmt.Errorf("failed to create TLS config: %v", err)
+	}
+
 	client := &http.Client{
 		Transport: &http.Transport{
-			TLSClientConfig: &tls.Config{
-				InsecureSkipVerify: ds.HTTPJson.TLS.SkipTlsVerify,
-			},
+			TLSClientConfig: tlsConfig,
 		},
 	}
 

center/router/router_event_pipeline.go

@@ -1,14 +1,19 @@
 package router
 
 import (
+	"encoding/json"
+	"fmt"
 	"net/http"
 	"time"
 
+	"github.com/ccfos/nightingale/v6/alert/pipeline/engine"
 	"github.com/ccfos/nightingale/v6/models"
 
 	"github.com/gin-gonic/gin"
+	"github.com/google/uuid"
 	"github.com/toolkits/pkg/ginx"
 	"github.com/toolkits/pkg/i18n"
+	"github.com/toolkits/pkg/logger"
 )
 
 // 获取事件Pipeline列表
@@ -27,6 +32,8 @@ func (rt *Router) eventPipelinesList(c *gin.Context) {
 		for _, tid := range pipeline.TeamIds {
 			pipeline.TeamNames = append(pipeline.TeamNames, ugMap[tid])
 		}
+		// 兼容处理：自动填充工作流字段
+		pipeline.FillWorkflowFields()
 	}
 
 	gids, err := models.MyGroupIdsMap(rt.Ctx, me.Id)
@@ -61,6 +68,9 @@ func (rt *Router) getEventPipeline(c *gin.Context) {
 	err = pipeline.FillTeamNames(rt.Ctx)
 	ginx.Dangerous(err)
 
+	// 兼容处理：自动填充工作流字段
+	pipeline.FillWorkflowFields()
+
 	ginx.NewRender(c).Data(pipeline, nil)
 }
 
@@ -131,7 +141,9 @@ func (rt *Router) tryRunEventPipeline(c *gin.Context) {
 	var f struct {
 		EventId        int64                `json:"event_id"`
 		PipelineConfig models.EventPipeline `json:"pipeline_config"`
+		EnvVariables   map[string]string    `json:"env_variables,omitempty"`
 	}
+
 	ginx.BindJSON(c, &f)
 
 	hisEvent, err := models.AlertHisEventGetById(rt.Ctx, f.EventId)
@@ -141,30 +153,33 @@ func (rt *Router) tryRunEventPipeline(c *gin.Context) {
 	event := hisEvent.ToCur()
 
 	lang := c.GetHeader("X-Language")
-	var result string
-	for _, p := range f.PipelineConfig.ProcessorConfigs {
-		processor, err := models.GetProcessorByType(p.Typ, p.Config)
-		if err != nil {
-			ginx.Bomb(http.StatusBadRequest, "get processor: %+v err: %+v", p, err)
-		}
-		event, result, err = processor.Process(rt.Ctx, event)
-		if err != nil {
-			ginx.Bomb(http.StatusBadRequest, "processor: %+v err: %+v", p, err)
-		}
+	me := c.MustGet("user").(*models.User)
 
-		if event == nil {
-			ginx.NewRender(c).Data(map[string]interface{}{
-				"event":  event,
-				"result": i18n.Sprintf(lang, "event is dropped"),
-			}, nil)
-			return
-		}
+	// 统一使用工作流引擎执行（兼容线性模式和工作流模式）
+	workflowEngine := engine.NewWorkflowEngine(rt.Ctx)
+
+	triggerCtx := &models.WorkflowTriggerContext{
+		Mode:         models.TriggerModeAPI,
+		TriggerBy:    me.Username,
+		EnvOverrides: f.EnvVariables,
+	}
+
+	resultEvent, result, err := workflowEngine.Execute(&f.PipelineConfig, event, triggerCtx)
+	if err != nil {
+		ginx.Bomb(http.StatusBadRequest, "pipeline execute error: %v", err)
 	}
 
 	m := map[string]interface{}{
-		"event":  event,
-		"result": i18n.Sprintf(lang, result),
+		"event":        resultEvent,
+		"result":       i18n.Sprintf(lang, result.Message),
+		"status":       result.Status,
+		"node_results": result.NodeResults,
 	}
+
+	if resultEvent == nil {
+		m["result"] = i18n.Sprintf(lang, "event is dropped")
+	}
+
 	ginx.NewRender(c).Data(m, nil)
 }
 
@@ -186,14 +201,18 @@ func (rt *Router) tryRunEventProcessor(c *gin.Context) {
 	if err != nil {
 		ginx.Bomb(200, "get processor err: %+v", err)
 	}
-	event, res, err := processor.Process(rt.Ctx, event)
+	wfCtx := &models.WorkflowContext{
+		Event: event,
+		Vars:  make(map[string]interface{}),
+	}
+	wfCtx, res, err := processor.Process(rt.Ctx, wfCtx)
 	if err != nil {
 		ginx.Bomb(200, "processor err: %+v", err)
 	}
 
 	lang := c.GetHeader("X-Language")
 	ginx.NewRender(c).Data(map[string]interface{}{
-		"event":  event,
+		"event":  wfCtx.Event,
 		"result": i18n.Sprintf(lang, res),
 	}, nil)
 }
@@ -223,6 +242,10 @@ func (rt *Router) tryRunEventProcessorByNotifyRule(c *gin.Context) {
 		ginx.Bomb(http.StatusBadRequest, "processors not found")
 	}
 
+	wfCtx := &models.WorkflowContext{
+		Event: event,
+		Vars:  make(map[string]interface{}),
+	}
 	for _, pl := range pipelines {
 		for _, p := range pl.ProcessorConfigs {
 			processor, err := models.GetProcessorByType(p.Typ, p.Config)
@@ -230,14 +253,14 @@ func (rt *Router) tryRunEventProcessorByNotifyRule(c *gin.Context) {
 				ginx.Bomb(http.StatusBadRequest, "get processor: %+v err: %+v", p, err)
 			}
 
-			event, _, err := processor.Process(rt.Ctx, event)
+			wfCtx, _, err = processor.Process(rt.Ctx, wfCtx)
 			if err != nil {
 				ginx.Bomb(http.StatusBadRequest, "processor: %+v err: %+v", p, err)
 			}
-			if event == nil {
+			if wfCtx == nil || wfCtx.Event == nil {
 				lang := c.GetHeader("X-Language")
 				ginx.NewRender(c).Data(map[string]interface{}{
-					"event":  event,
+					"event":  nil,
 					"result": i18n.Sprintf(lang, "event is dropped"),
 				}, nil)
 				return
@@ -245,10 +268,337 @@ func (rt *Router) tryRunEventProcessorByNotifyRule(c *gin.Context) {
 		}
 	}
 
-	ginx.NewRender(c).Data(event, nil)
+	ginx.NewRender(c).Data(wfCtx.Event, nil)
 }
 
 func (rt *Router) eventPipelinesListByService(c *gin.Context) {
 	pipelines, err := models.ListEventPipelines(rt.Ctx)
 	ginx.NewRender(c).Data(pipelines, err)
 }
+
+type EventPipelineRequest struct {
+	// 事件数据（可选，如果不传则使用空事件）
+	Event *models.AlertCurEvent `json:"event,omitempty"`
+	// 环境变量覆盖
+	EnvOverrides map[string]string `json:"env_overrides,omitempty"`
+
+	Username string `json:"username,omitempty"`
+}
+
+// executePipelineTrigger 执行 Pipeline 触发的公共逻辑
+func (rt *Router) executePipelineTrigger(pipeline *models.EventPipeline, req *EventPipelineRequest, triggerBy string) (string, error) {
+	// 准备事件数据
+	var event *models.AlertCurEvent
+	if req.Event != nil {
+		event = req.Event
+	} else {
+		// 创建空事件
+		event = &models.AlertCurEvent{
+			TriggerTime: time.Now().Unix(),
+		}
+	}
+
+	// 校验必填环境变量
+	if err := pipeline.ValidateEnvVariables(req.EnvOverrides); err != nil {
+		return "", fmt.Errorf("env validation failed: %v", err)
+	}
+
+	// 生成执行ID
+	executionID := uuid.New().String()
+
+	// 创建触发上下文
+	triggerCtx := &models.WorkflowTriggerContext{
+		Mode:         models.TriggerModeAPI,
+		TriggerBy:    triggerBy,
+		EnvOverrides: req.EnvOverrides,
+		RequestID:    executionID,
+	}
+
+	// 异步执行工作流
+	go func() {
+		workflowEngine := engine.NewWorkflowEngine(rt.Ctx)
+		_, _, err := workflowEngine.Execute(pipeline, event, triggerCtx)
+		if err != nil {
+			logger.Errorf("async workflow execute error: pipeline_id=%d execution_id=%s err=%v",
+				pipeline.ID, executionID, err)
+		}
+	}()
+
+	return executionID, nil
+}
+
+// triggerEventPipelineByService Service 调用触发工作流执行
+func (rt *Router) triggerEventPipelineByService(c *gin.Context) {
+	pipelineID := ginx.UrlParamInt64(c, "id")
+	var f EventPipelineRequest
+	ginx.BindJSON(c, &f)
+
+	// 获取 Pipeline
+	pipeline, err := models.GetEventPipeline(rt.Ctx, pipelineID)
+	if err != nil {
+		ginx.Bomb(http.StatusNotFound, "pipeline not found: %v", err)
+	}
+
+	executionID, err := rt.executePipelineTrigger(pipeline, &f, f.Username)
+	if err != nil {
+		ginx.Bomb(http.StatusBadRequest, "%v", err)
+	}
+
+	ginx.NewRender(c).Data(gin.H{
+		"execution_id": executionID,
+		"message":      "workflow execution started",
+	}, nil)
+}
+
+// triggerEventPipelineByAPI API 触发工作流执行
+func (rt *Router) triggerEventPipelineByAPI(c *gin.Context) {
+	pipelineID := ginx.UrlParamInt64(c, "id")
+	var f EventPipelineRequest
+	ginx.BindJSON(c, &f)
+
+	// 获取 Pipeline
+	pipeline, err := models.GetEventPipeline(rt.Ctx, pipelineID)
+	if err != nil {
+		ginx.Bomb(http.StatusNotFound, "pipeline not found: %v", err)
+	}
+
+	// 检查权限
+	me := c.MustGet("user").(*models.User)
+	ginx.Dangerous(me.CheckGroupPermission(rt.Ctx, pipeline.TeamIds))
+
+	executionID, err := rt.executePipelineTrigger(pipeline, &f, me.Username)
+	if err != nil {
+		ginx.Bomb(http.StatusBadRequest, err.Error())
+	}
+
+	ginx.NewRender(c).Data(gin.H{
+		"execution_id": executionID,
+		"message":      "workflow execution started",
+	}, nil)
+}
+
+func (rt *Router) listAllEventPipelineExecutions(c *gin.Context) {
+	pipelineName := ginx.QueryStr(c, "pipeline_name", "")
+	mode := ginx.QueryStr(c, "mode", "")
+	status := ginx.QueryStr(c, "status", "")
+	limit := ginx.QueryInt(c, "limit", 20)
+	offset := ginx.QueryInt(c, "p", 1)
+
+	if limit <= 0 || limit > 1000 {
+		limit = 20
+	}
+	if offset <= 0 {
+		offset = 1
+	}
+
+	executions, total, err := models.ListAllEventPipelineExecutions(rt.Ctx, pipelineName, mode, status, limit, (offset-1)*limit)
+	ginx.Dangerous(err)
+
+	ginx.NewRender(c).Data(gin.H{
+		"list":  executions,
+		"total": total,
+	}, nil)
+}
+
+func (rt *Router) listEventPipelineExecutions(c *gin.Context) {
+	pipelineID := ginx.UrlParamInt64(c, "id")
+	mode := ginx.QueryStr(c, "mode", "")
+	status := ginx.QueryStr(c, "status", "")
+	limit := ginx.QueryInt(c, "limit", 20)
+	offset := ginx.QueryInt(c, "p", 1)
+
+	if limit <= 0 || limit > 1000 {
+		limit = 20
+	}
+	if offset <= 0 {
+		offset = 1
+	}
+
+	executions, total, err := models.ListEventPipelineExecutions(rt.Ctx, pipelineID, mode, status, limit, (offset-1)*limit)
+	ginx.Dangerous(err)
+
+	ginx.NewRender(c).Data(gin.H{
+		"list":  executions,
+		"total": total,
+	}, nil)
+}
+
+func (rt *Router) getEventPipelineExecution(c *gin.Context) {
+	execID := ginx.UrlParamStr(c, "exec_id")
+
+	detail, err := models.GetEventPipelineExecutionDetail(rt.Ctx, execID)
+	if err != nil {
+		ginx.Bomb(http.StatusNotFound, "execution not found: %v", err)
+	}
+
+	ginx.NewRender(c).Data(detail, nil)
+}
+
+func (rt *Router) getEventPipelineExecutionStats(c *gin.Context) {
+	pipelineID := ginx.UrlParamInt64(c, "id")
+
+	stats, err := models.GetEventPipelineExecutionStatistics(rt.Ctx, pipelineID)
+	ginx.Dangerous(err)
+
+	ginx.NewRender(c).Data(stats, nil)
+}
+
+func (rt *Router) cleanEventPipelineExecutions(c *gin.Context) {
+	var f struct {
+		BeforeDays int `json:"before_days"`
+	}
+	ginx.BindJSON(c, &f)
+
+	if f.BeforeDays <= 0 {
+		f.BeforeDays = 30
+	}
+
+	beforeTime := time.Now().AddDate(0, 0, -f.BeforeDays).Unix()
+	affected, err := models.DeleteEventPipelineExecutions(rt.Ctx, beforeTime)
+	ginx.Dangerous(err)
+
+	ginx.NewRender(c).Data(gin.H{
+		"deleted": affected,
+	}, nil)
+}
+
+func (rt *Router) streamEventPipeline(c *gin.Context) {
+	pipelineID := ginx.UrlParamInt64(c, "id")
+
+	var f EventPipelineRequest
+	ginx.BindJSON(c, &f)
+
+	pipeline, err := models.GetEventPipeline(rt.Ctx, pipelineID)
+	if err != nil {
+		ginx.Bomb(http.StatusNotFound, "pipeline not found: %v", err)
+	}
+
+	me := c.MustGet("user").(*models.User)
+	ginx.Dangerous(me.CheckGroupPermission(rt.Ctx, pipeline.TeamIds))
+
+	var event *models.AlertCurEvent
+	if f.Event != nil {
+		event = f.Event
+	} else {
+		event = &models.AlertCurEvent{
+			TriggerTime: time.Now().Unix(),
+		}
+	}
+
+	triggerCtx := &models.WorkflowTriggerContext{
+		Mode:         models.TriggerModeAPI,
+		TriggerBy:    me.Username,
+		EnvOverrides: f.EnvOverrides,
+		RequestID:    uuid.New().String(),
+		Stream:       true, // 流式端点强制启用流式输出
+	}
+
+	workflowEngine := engine.NewWorkflowEngine(rt.Ctx)
+	_, result, err := workflowEngine.Execute(pipeline, event, triggerCtx)
+	if err != nil {
+		ginx.Bomb(http.StatusInternalServerError, "execute failed: %v", err)
+	}
+
+	if result.Stream && result.StreamChan != nil {
+		rt.handleStreamResponse(c, result, triggerCtx.RequestID)
+		return
+	}
+
+	ginx.NewRender(c).Data(result, nil)
+}
+
+func (rt *Router) handleStreamResponse(c *gin.Context, result *models.WorkflowResult, requestID string) {
+	// 设置 SSE 响应头
+	c.Header("Content-Type", "text/event-stream")
+	c.Header("Cache-Control", "no-cache")
+	c.Header("Connection", "keep-alive")
+	c.Header("X-Accel-Buffering", "no") // 禁用 nginx 缓冲
+	c.Header("X-Request-ID", requestID)
+
+	flusher, ok := c.Writer.(http.Flusher)
+	if !ok {
+		ginx.Bomb(http.StatusInternalServerError, "streaming not supported")
+		return
+	}
+
+	// 发送初始连接成功消息
+	initData := fmt.Sprintf(`{"type":"connected","request_id":"%s","timestamp":%d}`, requestID, time.Now().UnixMilli())
+	fmt.Fprintf(c.Writer, "data: %s\n\n", initData)
+	flusher.Flush()
+
+	// 从 channel 读取并发送 SSE
+	timeout := time.After(30 * time.Minute) // 最长流式输出时间
+	for {
+		select {
+		case chunk, ok := <-result.StreamChan:
+			if !ok {
+				// channel 关闭，发送结束标记
+				return
+			}
+
+			data, err := json.Marshal(chunk)
+			if err != nil {
+				logger.Errorf("stream: failed to marshal chunk: %v", err)
+				continue
+			}
+
+			fmt.Fprintf(c.Writer, "data: %s\n\n", data)
+			flusher.Flush()
+
+			if chunk.Done {
+				return
+			}
+
+		case <-c.Request.Context().Done():
+			// 客户端断开连接
+			logger.Infof("stream: client disconnected, request_id=%s", requestID)
+			return
+		case <-timeout:
+			logger.Errorf("stream: timeout, request_id=%s", requestID)
+			return
+		}
+	}
+}
+
+func (rt *Router) streamEventPipelineByService(c *gin.Context) {
+	pipelineID := ginx.UrlParamInt64(c, "id")
+
+	var f EventPipelineRequest
+	ginx.BindJSON(c, &f)
+
+	pipeline, err := models.GetEventPipeline(rt.Ctx, pipelineID)
+	if err != nil {
+		ginx.Bomb(http.StatusNotFound, "pipeline not found: %v", err)
+	}
+
+	var event *models.AlertCurEvent
+	if f.Event != nil {
+		event = f.Event
+	} else {
+		event = &models.AlertCurEvent{
+			TriggerTime: time.Now().Unix(),
+		}
+	}
+
+	triggerCtx := &models.WorkflowTriggerContext{
+		Mode:         models.TriggerModeAPI,
+		TriggerBy:    f.Username,
+		EnvOverrides: f.EnvOverrides,
+		RequestID:    uuid.New().String(),
+		Stream:       true, // 流式端点强制启用流式输出
+	}
+
+	workflowEngine := engine.NewWorkflowEngine(rt.Ctx)
+	_, result, err := workflowEngine.Execute(pipeline, event, triggerCtx)
+	if err != nil {
+		ginx.Bomb(http.StatusInternalServerError, "execute failed: %v", err)
+	}
+
+	// 检查是否是流式输出
+	if result.Stream && result.StreamChan != nil {
+		rt.handleStreamResponse(c, result, triggerCtx.RequestID)
+		return
+	}
+
+	ginx.NewRender(c).Data(result, nil)
+}

center/router/router_proxy.go

@@ -2,7 +2,6 @@ package router
 
 import (
 	"context"
-	"crypto/tls"
 	"fmt"
 	"net"
 	"net/http"
@@ -169,8 +168,15 @@ func (rt *Router) dsProxy(c *gin.Context) {
 
 	transport, has := transportGet(dsId, ds.UpdatedAt)
 	if !has {
+		// 使用 TLS 配置（支持 mTLS）
+		tlsConfig, err := ds.HTTPJson.TLS.TLSConfig()
+		if err != nil {
+			c.String(http.StatusInternalServerError, "failed to create TLS config: %s", err.Error())
+			return
+		}
+
 		transport = &http.Transport{
-			TLSClientConfig: &tls.Config{InsecureSkipVerify: ds.HTTPJson.TLS.SkipTlsVerify},
+			TLSClientConfig: tlsConfig,
 			Proxy:           http.ProxyFromEnvironment,
 			DialContext: (&net.Dialer{
 				Timeout: time.Duration(ds.HTTPJson.DialTimeout) * time.Millisecond,

center/router/router_query.go

@@ -5,6 +5,7 @@ import (
 	"sort"
 	"sync"
 
+	"github.com/ccfos/nightingale/v6/alert/eval"
 	"github.com/ccfos/nightingale/v6/dscache"
 	"github.com/ccfos/nightingale/v6/models"
 	"github.com/gin-gonic/gin"
@@ -12,7 +13,9 @@ import (
 	"github.com/toolkits/pkg/logger"
 )
 
-func CheckDsPerm(c *gin.Context, dsId int64, cate string, q interface{}) bool {
+type CheckDsPermFunc func(c *gin.Context, dsId int64, cate string, q interface{}) bool
+
+var CheckDsPerm CheckDsPermFunc = func(c *gin.Context, dsId int64, cate string, q interface{}) bool {
 	// todo: 后续需要根据 cate 判断是否需要权限
 	return true
 }
@@ -56,6 +59,13 @@ func QueryLogBatchConcurrently(anonymousAccess bool, ctx *gin.Context, f QueryFr
 			return LogResp{}, fmt.Errorf("cluster not exists")
 		}
 
+		// 根据数据源类型对 Query 进行模板渲染处理
+		err := eval.ExecuteQueryTemplate(q.DsCate, q.Query, nil)
+		if err != nil {
+			logger.Warningf("query template execute error: %v", err)
+			return LogResp{}, fmt.Errorf("query template execute error: %v", err)
+		}
+
 		wg.Add(1)
 		go func(query Query) {
 			defer wg.Done()

center/router/router_tdengine.go

@@ -2,13 +2,14 @@ package router
 
 import (
 	"fmt"
+	"net/http"
+
 	"github.com/ccfos/nightingale/v6/center/cconf"
 	"github.com/ccfos/nightingale/v6/datasource/tdengine"
 	"github.com/ccfos/nightingale/v6/dscache"
 	"github.com/ccfos/nightingale/v6/models"
 	"github.com/gin-gonic/gin"
 	"github.com/toolkits/pkg/ginx"
-	"net/http"
 )
 
 type databasesQueryForm struct {

docker/compose-bridge/docker-compose.yaml

@@ -87,8 +87,8 @@ services:
       - mysql
       - redis
       - victoriametrics
-    command: >
-      sh -c "/app/n9e"
+    command:
+      - /app/n9e
 
   categraf:
     image: "flashcatcloud/categraf:latest"

docker/compose-host-network-metric-log/docker-compose.yaml

@@ -59,8 +59,8 @@ services:
       - mysql
       - redis
       - prometheus
-    command: >
-      sh -c "/app/n9e"
+    command:
+      - /app/n9e
 
   categraf:
     image: "flashcatcloud/categraf:latest"

docker/compose-host-network/docker-compose.yaml

@@ -58,8 +58,8 @@ services:
       - mysql
       - redis
       - prometheus
-    command: >
-      sh -c "/app/n9e"
+    command:
+      - /app/n9e
 
   categraf:
     image: "flashcatcloud/categraf:latest"

docker/compose-postgres/docker-compose.yaml

@@ -74,8 +74,8 @@ services:
       - postgres:postgres
       - redis:redis
       - victoriametrics:victoriametrics
-    command: >
-      sh -c "/app/n9e"
+    command:
+      - /app/n9e
 
   categraf:
     image: "flashcatcloud/categraf:latest"

docker/compose-postgres/initsql_for_postgres/a-n9e-for-Postgres.sql

@@ -804,6 +804,9 @@ CREATE TABLE builtin_metrics (
   lang varchar(191) NOT NULL DEFAULT '',
   note varchar(4096) NOT NULL,
   expression varchar(4096) NOT NULL,
+  expression_type varchar(32) NOT NULL DEFAULT 'promql',
+  metric_type varchar(191) NOT NULL DEFAULT '',
+  extra_fields text,
   created_at bigint NOT NULL DEFAULT 0,
   created_by varchar(191) NOT NULL DEFAULT '',
   updated_at bigint NOT NULL DEFAULT 0,
@@ -826,6 +829,9 @@ COMMENT ON COLUMN builtin_metrics.unit IS 'unit of metric';
 COMMENT ON COLUMN builtin_metrics.lang IS 'language of metric';
 COMMENT ON COLUMN builtin_metrics.note IS 'description of metric in Chinese';
 COMMENT ON COLUMN builtin_metrics.expression IS 'expression of metric';
+COMMENT ON COLUMN builtin_metrics.expression_type IS 'expression type: metric_name or promql';
+COMMENT ON COLUMN builtin_metrics.metric_type IS 'metric type like counter/gauge';
+COMMENT ON COLUMN builtin_metrics.extra_fields IS 'custom extra fields';
 COMMENT ON COLUMN builtin_metrics.created_at IS 'create time';
 COMMENT ON COLUMN builtin_metrics.created_by IS 'creator';
 COMMENT ON COLUMN builtin_metrics.updated_at IS 'update time';

docker/initsql/a-n9e.sql

@@ -719,6 +719,9 @@ CREATE TABLE `builtin_metrics` (
     `lang` varchar(191) NOT NULL DEFAULT 'zh' COMMENT '''language''',
     `note` varchar(4096) NOT NULL COMMENT '''description of metric''',
     `expression` varchar(4096) NOT NULL COMMENT '''expression of metric''',
+    `expression_type` varchar(32) NOT NULL DEFAULT 'promql' COMMENT '''expression type: metric_name or promql''',
+    `metric_type` varchar(191) NOT NULL DEFAULT '' COMMENT '''metric type like counter/gauge''',
+    `extra_fields` text COMMENT '''custom extra fields''',
     `created_at` bigint NOT NULL DEFAULT 0 COMMENT '''create time''',
     `created_by` varchar(191) NOT NULL DEFAULT '' COMMENT '''creator''',
     `updated_at` bigint NOT NULL DEFAULT 0 COMMENT '''update time''',

docker/migratesql/migrate.sql

@@ -292,4 +292,42 @@ ALTER TABLE `alert_rule` ADD COLUMN `pipeline_configs` text COMMENT 'pipeline co
 
 /* v8.4.2 2025-11-13 */
 ALTER TABLE `board` ADD COLUMN `note` varchar(1024) not null default '' comment 'note';
-ALTER TABLE `builtin_payloads` ADD COLUMN `note` varchar(1024) not null default '' comment 'note of payload';
+ALTER TABLE `builtin_payloads` ADD COLUMN `note` varchar(1024) not null default '' comment 'note of payload';
+
+/* v9 2026-01-09 */
+ALTER TABLE `event_pipeline` ADD COLUMN `typ` varchar(128) NOT NULL DEFAULT '' COMMENT 'pipeline type: builtin, user-defined';
+ALTER TABLE `event_pipeline` ADD COLUMN `use_case` varchar(128) NOT NULL DEFAULT '' COMMENT 'use case: metric_explorer, event_summary, event_pipeline';
+ALTER TABLE `event_pipeline` ADD COLUMN `trigger_mode` varchar(128) NOT NULL DEFAULT 'event' COMMENT 'trigger mode: event, api, cron';
+ALTER TABLE `event_pipeline` ADD COLUMN `disabled` tinyint(1) NOT NULL DEFAULT 0 COMMENT 'disabled flag';
+ALTER TABLE `event_pipeline` ADD COLUMN `nodes` text COMMENT 'workflow nodes (JSON)';
+ALTER TABLE `event_pipeline` ADD COLUMN `connections` text COMMENT 'node connections (JSON)';
+ALTER TABLE `event_pipeline` ADD COLUMN `env_variables` text COMMENT 'environment variables (JSON)';
+ALTER TABLE `event_pipeline` ADD COLUMN `label_filters` text COMMENT 'label filters (JSON)';
+
+CREATE TABLE `event_pipeline_execution` (
+    `id` varchar(36) NOT NULL COMMENT 'execution id',
+    `pipeline_id` bigint NOT NULL COMMENT 'pipeline id',
+    `pipeline_name` varchar(128) DEFAULT '' COMMENT 'pipeline name snapshot',
+    `event_id` bigint DEFAULT 0 COMMENT 'related alert event id',
+    `mode` varchar(16) NOT NULL DEFAULT 'event' COMMENT 'trigger mode: event/api/cron',
+    `status` varchar(16) NOT NULL DEFAULT 'running' COMMENT 'status: running/success/failed',
+    `node_results` mediumtext COMMENT 'node execution results (JSON)',
+    `error_message` varchar(1024) DEFAULT '' COMMENT 'error message',
+    `error_node` varchar(36) DEFAULT '' COMMENT 'error node id',
+    `created_at` bigint NOT NULL DEFAULT 0 COMMENT 'start timestamp',
+    `finished_at` bigint DEFAULT 0 COMMENT 'finish timestamp',
+    `duration_ms` bigint DEFAULT 0 COMMENT 'duration in milliseconds',
+    `trigger_by` varchar(64) DEFAULT '' COMMENT 'trigger by',
+    `env_snapshot` text COMMENT 'environment variables snapshot (sanitized)',
+    PRIMARY KEY (`id`),
+    KEY `idx_pipeline_id` (`pipeline_id`),
+    KEY `idx_event_id` (`event_id`),
+    KEY `idx_mode` (`mode`),
+    KEY `idx_status` (`status`),
+    KEY `idx_created_at` (`created_at`)
+) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COMMENT='event pipeline execution records';
+
+/* v8.5.0 builtin_metrics new fields */
+ALTER TABLE `builtin_metrics` ADD COLUMN `expression_type` varchar(32) NOT NULL DEFAULT 'promql' COMMENT 'expression type: metric_name or promql';
+ALTER TABLE `builtin_metrics` ADD COLUMN `metric_type` varchar(191) NOT NULL DEFAULT '' COMMENT 'metric type like counter/gauge';
+ALTER TABLE `builtin_metrics` ADD COLUMN `extra_fields` text COMMENT 'custom extra fields';

docker/sqlite.sql

@@ -651,6 +651,9 @@ CREATE TABLE `builtin_metrics` (
     `lang` varchar(191) NOT NULL DEFAULT '',
     `note` varchar(4096) NOT NULL,
     `expression` varchar(4096) NOT NULL,
+    `expression_type` varchar(32) NOT NULL DEFAULT 'promql',
+    `metric_type` varchar(191) NOT NULL DEFAULT '',
+    `extra_fields` text,
     `created_at` bigint NOT NULL DEFAULT 0,
     `created_by` varchar(191) NOT NULL DEFAULT '',
     `updated_at` bigint NOT NULL DEFAULT 0,

dscache/cache.go

@@ -57,3 +57,29 @@ func (cs *Cache) Get(cate string, dsId int64) (datasource.Datasource, bool) {
 
 	return cs.datas[cate][dsId], true
 }
+
+func (cs *Cache) Delete(cate string, dsId int64) {
+	cs.mutex.Lock()
+	defer cs.mutex.Unlock()
+	if _, found := cs.datas[cate]; !found {
+		return
+	}
+	delete(cs.datas[cate], dsId)
+
+	logger.Debugf("delete plugin:%s %d from cache", cate, dsId)
+}
+
+// GetAllIds 返回缓存中所有数据源的 ID，按类型分组
+func (cs *Cache) GetAllIds() map[string][]int64 {
+	cs.mutex.RLock()
+	defer cs.mutex.RUnlock()
+	result := make(map[string][]int64)
+	for cate, dsMap := range cs.datas {
+		ids := make([]int64, 0, len(dsMap))
+		for dsId := range dsMap {
+			ids = append(ids, dsId)
+		}
+		result[cate] = ids
+	}
+	return result
+}

dscache/sync.go

@@ -173,7 +173,10 @@ func esN9eToDatasourceInfo(ds *datasource.DatasourceInfo, item models.Datasource
 }
 
 func PutDatasources(items []datasource.DatasourceInfo) {
+	// 记录当前有效的数据源 ID，按类型分组
+	validIds := make(map[string]map[int64]struct{})
 	ids := make([]int64, 0)
+
 	for _, item := range items {
 		if item.Type == "prometheus" {
 			continue
@@ -202,6 +205,12 @@ func PutDatasources(items []datasource.DatasourceInfo) {
 		}
 		ids = append(ids, item.Id)
 
+		// 记录有效的数据源 ID
+		if _, ok := validIds[typ]; !ok {
+			validIds[typ] = make(map[int64]struct{})
+		}
+		validIds[typ][item.Id] = struct{}{}
+
 		// 异步初始化 client 不然数据源同步的会很慢
 		go func() {
 			defer func() {
@@ -213,5 +222,19 @@ func PutDatasources(items []datasource.DatasourceInfo) {
 		}()
 	}
 
+	// 删除 items 中不存在但 DsCache 中存在的数据源
+	cachedIds := DsCache.GetAllIds()
+	for cate, dsIds := range cachedIds {
+		for _, dsId := range dsIds {
+			if _, ok := validIds[cate]; !ok {
+				// 该类型在 items 中完全不存在，删除缓存中的所有该类型数据源
+				DsCache.Delete(cate, dsId)
+			} else if _, ok := validIds[cate][dsId]; !ok {
+				// 该数据源 ID 在 items 中不存在，删除
+				DsCache.Delete(cate, dsId)
+			}
+		}
+	}
+
 	logger.Debugf("get plugin by type success Ids:%v", ids)
 }

dskit/doris/doris.go

@@ -32,7 +32,7 @@ type Doris struct {
 	FeAddr          string `json:"doris.fe_addr" mapstructure:"doris.fe_addr"`   // fe http endpoint
 	User            string `json:"doris.user" mapstructure:"doris.user"`         //
 	Password        string `json:"doris.password" mapstructure:"doris.password"` //
-	Timeout         int    `json:"doris.timeout" mapstructure:"doris.timeout"`
+	Timeout         int    `json:"doris.timeout" mapstructure:"doris.timeout"`   // ms
 	MaxIdleConns    int    `json:"doris.max_idle_conns" mapstructure:"doris.max_idle_conns"`
 	MaxOpenConns    int    `json:"doris.max_open_conns" mapstructure:"doris.max_open_conns"`
 	ConnMaxLifetime int    `json:"doris.conn_max_lifetime" mapstructure:"doris.conn_max_lifetime"`
@@ -71,7 +71,7 @@ func (d *Doris) NewConn(ctx context.Context, database string) (*sql.DB, error) {
 
 	// Set default values similar to postgres implementation
 	if d.Timeout == 0 {
-		d.Timeout = 60
+		d.Timeout = 60000
 	}
 	if d.MaxIdleConns == 0 {
 		d.MaxIdleConns = 10
@@ -127,7 +127,7 @@ func (d *Doris) createTimeoutContext(ctx context.Context) (context.Context, cont
 	if timeout == 0 {
 		timeout = 60
 	}
-	return context.WithTimeout(ctx, time.Duration(timeout)*time.Second)
+	return context.WithTimeout(ctx, time.Duration(timeout)*time.Millisecond)
 }
 
 // ShowDatabases lists all databases in Doris

integrations/Java/dashboards/jmx_by_kubernetes.json

@@ -1,5 +1,4 @@
-[
-  {
+{
     "name": "JMX - Kubernetes",
     "tags": "Prometheus JMX Kubernetes",
     "configs": {
@@ -1871,5 +1870,4 @@
       "version": "3.0.0"
     },
     "uuid": 1755595969673000
-  }
-]
+  }

memsto/config_cval_cache.go

@@ -125,6 +125,7 @@ type SiteInfo struct {
 	PrintBodyPaths []string `json:"print_body_paths"`
 	PrintAccessLog bool     `json:"print_access_log"`
 	SiteUrl        string   `json:"site_url"`
+	ReportHostNIC  bool     `json:"report_host_nic"`
 }
 
 func (c *CvalCache) GetSiteInfo() *SiteInfo {

memsto/event_processor_cache.go

@@ -68,18 +68,6 @@ func (epc *EventProcessorCacheType) Get(processorId int64) *models.EventPipeline
 	return epc.eventPipelines[processorId]
 }
 
-func (epc *EventProcessorCacheType) GetProcessorsById(processorId int64) []models.Processor {
-	epc.RLock()
-	defer epc.RUnlock()
-
-	eventPipeline, ok := epc.eventPipelines[processorId]
-	if !ok {
-		return []models.Processor{}
-	}
-
-	return eventPipeline.Processors
-}
-
 func (epc *EventProcessorCacheType) GetProcessorIds() []int64 {
 	epc.RLock()
 	defer epc.RUnlock()
@@ -137,18 +125,7 @@ func (epc *EventProcessorCacheType) syncEventProcessors() error {
 
 	m := make(map[int64]*models.EventPipeline)
 	for i := 0; i < len(lst); i++ {
-		eventPipeline := lst[i]
-		for _, p := range eventPipeline.ProcessorConfigs {
-			processor, err := models.GetProcessorByType(p.Typ, p.Config)
-			if err != nil {
-				logger.Warningf("event_pipeline_id: %d, event:%+v, processor:%+v get processor err: %+v", eventPipeline.ID, eventPipeline, p, err)
-				continue
-			}
-
-			eventPipeline.Processors = append(eventPipeline.Processors, processor)
-		}
-
-		m[lst[i].ID] = eventPipeline
+		m[lst[i].ID] = lst[i]
 	}
 
 	epc.Set(m, stat.Total, stat.LastUpdated)

memsto/notify_channel_cache.go

@@ -283,7 +283,7 @@ func (ncc *NotifyChannelCacheType) processNotifyTask(task *NotifyTask) {
 		if len(task.Sendtos) == 0 || ncc.needBatchContacts(task.NotifyChannel.RequestConfig.HTTPRequestConfig) {
 			start := time.Now()
 			resp, err := task.NotifyChannel.SendHTTP(task.Events, task.TplContent, task.CustomParams, task.Sendtos, httpClient)
-			resp = fmt.Sprintf("duration: %d ms %s", time.Since(start).Milliseconds(), resp)
+			resp = fmt.Sprintf("send_time: %s duration: %d ms %s", time.Now().Format("2006-01-02 15:04:05"), time.Since(start).Milliseconds(), resp)
 			logger.Infof("http_sendernotify_id: %d, channel_name: %v, event:%+v, tplContent:%v, customParams:%v, userInfo:%+v, respBody: %v, err: %v",
 				task.NotifyRuleId, task.NotifyChannel.Name, task.Events[0], task.TplContent, task.CustomParams, task.Sendtos, resp, err)
 

models/builtin_metrics.go

@@ -1,6 +1,7 @@
 package models
 
 import (
+	"encoding/json"
 	"errors"
 	"fmt"
 	"strings"
@@ -12,20 +13,23 @@ import (
 
 // BuiltinMetric represents a metric along with its metadata.
 type BuiltinMetric struct {
-	ID          int64         `json:"id" gorm:"primaryKey;type:bigint;autoIncrement;comment:'unique identifier'"`
-	UUID        int64         `json:"uuid" gorm:"type:bigint;not null;default:0;comment:'uuid'"`
-	Collector   string        `json:"collector" gorm:"type:varchar(191);not null;index:idx_collector,sort:asc;comment:'type of collector'"`
-	Typ         string        `json:"typ" gorm:"type:varchar(191);not null;index:idx_typ,sort:asc;comment:'type of metric'"`
-	Name        string        `json:"name" gorm:"type:varchar(191);not null;index:idx_builtinmetric_name,sort:asc;comment:'name of metric'"`
-	Unit        string        `json:"unit" gorm:"type:varchar(191);not null;comment:'unit of metric'"`
-	Note        string        `json:"note" gorm:"type:varchar(4096);not null;comment:'description of metric'"`
-	Lang        string        `json:"lang" gorm:"type:varchar(191);not null;default:'zh';index:idx_lang,sort:asc;comment:'language'"`
-	Translation []Translation `json:"translation" gorm:"type:text;serializer:json;comment:'translation of metric'"`
-	Expression  string        `json:"expression" gorm:"type:varchar(4096);not null;comment:'expression of metric'"`
-	CreatedAt   int64         `json:"created_at" gorm:"type:bigint;not null;default:0;comment:'create time'"`
-	CreatedBy   string        `json:"created_by" gorm:"type:varchar(191);not null;default:'';comment:'creator'"`
-	UpdatedAt   int64         `json:"updated_at" gorm:"type:bigint;not null;default:0;comment:'update time'"`
-	UpdatedBy   string        `json:"updated_by" gorm:"type:varchar(191);not null;default:'';comment:'updater'"`
+	ID             int64           `json:"id" gorm:"primaryKey;type:bigint;autoIncrement;comment:'unique identifier'"`
+	UUID           int64           `json:"uuid" gorm:"type:bigint;not null;default:0;comment:'uuid'"`
+	Collector      string          `json:"collector" gorm:"type:varchar(191);not null;index:idx_collector,sort:asc;comment:'type of collector'"`
+	Typ            string          `json:"typ" gorm:"type:varchar(191);not null;index:idx_typ,sort:asc;comment:'type of metric'"`
+	Name           string          `json:"name" gorm:"type:varchar(191);not null;index:idx_builtinmetric_name,sort:asc;comment:'name of metric'"`
+	Unit           string          `json:"unit" gorm:"type:varchar(191);not null;comment:'unit of metric'"`
+	Note           string          `json:"note" gorm:"type:varchar(4096);not null;comment:'description of metric'"`
+	Lang           string          `json:"lang" gorm:"type:varchar(191);not null;default:'zh';index:idx_lang,sort:asc;comment:'language'"`
+	Translation    []Translation   `json:"translation" gorm:"type:text;serializer:json;comment:'translation of metric'"`
+	Expression     string          `json:"expression" gorm:"type:varchar(4096);not null;comment:'expression of metric'"`
+	ExpressionType string          `json:"expression_type" gorm:"type:varchar(32);not null;default:'promql';comment:'expression type: metric_name or promql'"`
+	MetricType     string          `json:"metric_type" gorm:"type:varchar(191);not null;default:'';comment:'metric type like counter/gauge'"`
+	ExtraFields    json.RawMessage `json:"extra_fields" gorm:"type:text;serializer:json;comment:'custom extra fields'"`
+	CreatedAt      int64           `json:"created_at" gorm:"type:bigint;not null;default:0;comment:'create time'"`
+	CreatedBy      string          `json:"created_by" gorm:"type:varchar(191);not null;default:'';comment:'creator'"`
+	UpdatedAt      int64           `json:"updated_at" gorm:"type:bigint;not null;default:0;comment:'update time'"`
+	UpdatedBy      string          `json:"updated_by" gorm:"type:varchar(191);not null;default:'';comment:'updater'"`
 }
 
 type Translation struct {

models/builtin_payload.go

@@ -29,6 +29,27 @@ func (bp *BuiltinPayload) TableName() string {
 	return "builtin_payloads"
 }
 
+type PostgresBuiltinPayload struct {
+	ID          int64  `json:"id" gorm:"primaryKey;type:bigint;autoIncrement;comment:'unique identifier'"`
+	Type        string `json:"type" gorm:"type:varchar(191);not null;index:idx_type,sort:asc;comment:'type of payload'"`
+	Component   string `json:"component" gorm:"type:varchar(191);not null;index:idx_component,sort:asc;comment:'component of payload'"`
+	ComponentID uint64 `json:"component_id" gorm:"type:bigint;index:idx_component,sort:asc;comment:'component_id of payload'"`
+	Cate        string `json:"cate" gorm:"type:varchar(191);not null;comment:'category of payload'"`
+	Name        string `json:"name" gorm:"type:varchar(191);not null;index:idx_buildinpayload_name,sort:asc;comment:'name of payload'"`
+	Tags        string `json:"tags" gorm:"type:varchar(191);not null;default:'';comment:'tags of payload'"`
+	Content     string `json:"content" gorm:"type:text;not null;comment:'content of payload'"`
+	UUID        int64  `json:"uuid" gorm:"type:bigint;not null;index:idx_uuid;comment:'uuid of payload'"`
+	Note        string `json:"note" gorm:"type:varchar(1024);not null;default:'';comment:'note of payload'"`
+	CreatedAt   int64  `json:"created_at" gorm:"type:bigint;not null;default:0;comment:'create time'"`
+	CreatedBy   string `json:"created_by" gorm:"type:varchar(191);not null;default:'';comment:'creator'"`
+	UpdatedAt   int64  `json:"updated_at" gorm:"type:bigint;not null;default:0;comment:'update time'"`
+	UpdatedBy   string `json:"updated_by" gorm:"type:varchar(191);not null;default:'';comment:'updater'"`
+}
+
+func (bp *PostgresBuiltinPayload) TableName() string {
+	return "builtin_payloads"
+}
+
 func (bp *BuiltinPayload) Verify() error {
 	bp.Type = strings.TrimSpace(bp.Type)
 	if bp.Type == "" {

models/datasource.go

@@ -1,7 +1,10 @@
 package models
 
 import (
+	"crypto/tls"
+	"crypto/x509"
 	"encoding/json"
+	"fmt"
 	"math/rand"
 	"net/http"
 	"net/url"
@@ -144,6 +147,72 @@ func (h HTTP) ParseUrl() (target *url.URL, err error) {
 
 type TLS struct {
 	SkipTlsVerify bool `json:"skip_tls_verify"`
+	// mTLS 配置
+	CACert            string `json:"ca_cert"`             // CA 证书内容 (PEM 格式)
+	ClientCert        string `json:"client_cert"`         // 客户端证书内容 (PEM 格式)
+	ClientKey         string `json:"client_key"`          // 客户端密钥内容 (PEM 格式)
+	ClientKeyPassword string `json:"client_key_password"` // 密钥密码（可选）
+	ServerName        string `json:"server_name"`         // TLS ServerName（可选，用于证书验证）
+	MinVersion        string `json:"min_version"`         // TLS 最小版本 (1.0, 1.1, 1.2, 1.3)
+	MaxVersion        string `json:"max_version"`         // TLS 最大版本
+}
+
+// TLSConfig 从证书内容创建 tls.Config
+// 证书内容为 PEM 格式字符串
+func (t *TLS) TLSConfig() (*tls.Config, error) {
+	tlsConfig := &tls.Config{
+		InsecureSkipVerify: t.SkipTlsVerify,
+	}
+
+	// 设置 ServerName
+	if t.ServerName != "" {
+		tlsConfig.ServerName = t.ServerName
+	}
+
+	// 设置 TLS 版本
+	if t.MinVersion != "" {
+		if v, ok := tlsVersionMap[t.MinVersion]; ok {
+			tlsConfig.MinVersion = v
+		}
+	}
+	if t.MaxVersion != "" {
+		if v, ok := tlsVersionMap[t.MaxVersion]; ok {
+			tlsConfig.MaxVersion = v
+		}
+	}
+
+	// 如果配置了客户端证书，则加载 mTLS 配置
+	clientCert := strings.TrimSpace(t.ClientCert)
+	clientKey := strings.TrimSpace(t.ClientKey)
+	caCert := strings.TrimSpace(t.CACert)
+
+	if clientCert != "" && clientKey != "" {
+		// 加载客户端证书和密钥
+		cert, err := tls.X509KeyPair([]byte(clientCert), []byte(clientKey))
+		if err != nil {
+			return nil, fmt.Errorf("failed to load client certificate: %w", err)
+		}
+		tlsConfig.Certificates = []tls.Certificate{cert}
+	}
+
+	// 加载 CA 证书
+	if caCert != "" {
+		caCertPool := x509.NewCertPool()
+		if !caCertPool.AppendCertsFromPEM([]byte(caCert)) {
+			return nil, fmt.Errorf("failed to parse CA certificate")
+		}
+		tlsConfig.RootCAs = caCertPool
+	}
+
+	return tlsConfig, nil
+}
+
+// tlsVersionMap TLS 版本映射
+var tlsVersionMap = map[string]uint16{
+	"1.0": tls.VersionTLS10,
+	"1.1": tls.VersionTLS11,
+	"1.2": tls.VersionTLS12,
+	"1.3": tls.VersionTLS13,
 }
 
 func (ds *Datasource) TableName() string {

models/event_pipeline.go

@@ -13,6 +13,10 @@ import (
 type EventPipeline struct {
 	ID               int64             `json:"id" gorm:"primaryKey"`
 	Name             string            `json:"name" gorm:"type:varchar(128)"`
+	Typ              string            `json:"typ" gorm:"type:varchar(128)"`          // builtin, user-defined    // event_pipeline, event_summary, metric_explorer
+	UseCase          string            `json:"use_case" gorm:"type:varchar(128)"`     // metric_explorer, event_summary, event_pipeline
+	TriggerMode      string            `json:"trigger_mode" gorm:"type:varchar(128)"` // event, api, cron
+	Disabled         bool              `json:"disabled" gorm:"type:boolean"`
 	TeamIds          []int64           `json:"team_ids" gorm:"type:text;serializer:json"`
 	TeamNames        []string          `json:"team_names" gorm:"-"`
 	Description      string            `json:"description" gorm:"type:varchar(255)"`
@@ -20,12 +24,18 @@ type EventPipeline struct {
 	LabelFilters     []TagFilter       `json:"label_filters" gorm:"type:text;serializer:json"`
 	AttrFilters      []TagFilter       `json:"attribute_filters" gorm:"type:text;serializer:json"`
 	ProcessorConfigs []ProcessorConfig `json:"processors" gorm:"type:text;serializer:json"`
-	CreateAt         int64             `json:"create_at" gorm:"type:bigint"`
-	CreateBy         string            `json:"create_by" gorm:"type:varchar(64)"`
-	UpdateAt         int64             `json:"update_at" gorm:"type:bigint"`
-	UpdateBy         string            `json:"update_by" gorm:"type:varchar(64)"`
 
-	Processors []Processor `json:"-" gorm:"-"`
+	// 工作流节点列表
+	Nodes []WorkflowNode `json:"nodes,omitempty" gorm:"type:text;serializer:json"`
+	// 节点连接关系
+	Connections Connections `json:"connections,omitempty" gorm:"type:text;serializer:json"`
+	// 环境变量（工作流级别的配置变量）
+	EnvVariables []EnvVariable `json:"env_variables,omitempty" gorm:"type:text;serializer:json"`
+
+	CreateAt int64  `json:"create_at" gorm:"type:bigint"`
+	CreateBy string `json:"create_by" gorm:"type:varchar(64)"`
+	UpdateAt int64  `json:"update_at" gorm:"type:bigint"`
+	UpdateBy string `json:"update_by" gorm:"type:varchar(64)"`
 }
 
 type ProcessorConfig struct {
@@ -46,9 +56,6 @@ func (e *EventPipeline) Verify() error {
 		return errors.New("team_ids cannot be empty")
 	}
 
-	if len(e.TeamIds) == 0 {
-		e.TeamIds = make([]int64, 0)
-	}
 	if len(e.LabelFilters) == 0 {
 		e.LabelFilters = make([]TagFilter, 0)
 	}
@@ -59,6 +66,17 @@ func (e *EventPipeline) Verify() error {
 		e.ProcessorConfigs = make([]ProcessorConfig, 0)
 	}
 
+	// 初始化空数组，避免 null
+	if e.Nodes == nil {
+		e.Nodes = make([]WorkflowNode, 0)
+	}
+	if e.Connections == nil {
+		e.Connections = make(Connections)
+	}
+	if e.EnvVariables == nil {
+		e.EnvVariables = make([]EnvVariable, 0)
+	}
+
 	return nil
 }
 
@@ -176,3 +194,87 @@ func EventPipelineStatistics(ctx *ctx.Context) (*Statistics, error) {
 
 	return stats[0], nil
 }
+
+// 无论是新格式还是旧格式，都返回统一的 []WorkflowNode
+func (e *EventPipeline) GetWorkflowNodes() []WorkflowNode {
+	// 优先使用新格式
+	if len(e.Nodes) > 0 {
+		return e.Nodes
+	}
+
+	// 兼容旧格式：将 ProcessorConfigs 转换为 WorkflowNode
+	nodes := make([]WorkflowNode, len(e.ProcessorConfigs))
+	for i, pc := range e.ProcessorConfigs {
+		nodeID := fmt.Sprintf("node_%d", i)
+		nodeName := pc.Typ
+
+		nodes[i] = WorkflowNode{
+			ID:     nodeID,
+			Name:   nodeName,
+			Type:   pc.Typ,
+			Config: pc.Config,
+		}
+	}
+	return nodes
+}
+
+func (e *EventPipeline) GetWorkflowConnections() Connections {
+	// 优先使用显式定义的连接
+	if len(e.Connections) > 0 {
+		return e.Connections
+	}
+
+	// 自动生成线性连接：node_0 → node_1 → node_2 → ...
+	nodes := e.GetWorkflowNodes()
+	conns := make(Connections)
+
+	for i := 0; i < len(nodes)-1; i++ {
+		conns[nodes[i].ID] = NodeConnections{
+			Main: [][]ConnectionTarget{
+				{{Node: nodes[i+1].ID, Type: "main", Index: 0}},
+			},
+		}
+	}
+	return conns
+}
+
+func (e *EventPipeline) FillWorkflowFields() {
+	if len(e.Nodes) == 0 && len(e.ProcessorConfigs) > 0 {
+		e.Nodes = e.GetWorkflowNodes()
+		e.Connections = e.GetWorkflowConnections()
+	}
+}
+
+func (e *EventPipeline) GetEnvMap() map[string]string {
+	envMap := make(map[string]string)
+	for _, v := range e.EnvVariables {
+		envMap[v.Key] = v.Value
+	}
+	return envMap
+}
+
+func (e *EventPipeline) GetSecretKeys() map[string]bool {
+	secretKeys := make(map[string]bool)
+	for _, v := range e.EnvVariables {
+		if v.Secret {
+			secretKeys[v.Key] = true
+		}
+	}
+	return secretKeys
+}
+
+func (e *EventPipeline) ValidateEnvVariables(overrides map[string]string) error {
+	// 合并默认值和覆盖值
+	merged := e.GetEnvMap()
+	for k, v := range overrides {
+		merged[k] = v
+	}
+
+	// 校验必填项
+	for _, v := range e.EnvVariables {
+		if v.Required && merged[v.Key] == "" {
+			return fmt.Errorf("required env variable %s is missing", v.Key)
+		}
+	}
+	return nil
+}

models/event_pipeline_execution.go

@@ -0,0 +1,282 @@
+package models
+
+import (
+	"encoding/json"
+	"fmt"
+
+	"github.com/ccfos/nightingale/v6/pkg/ctx"
+)
+
+// 执行状态常量
+const (
+	ExecutionStatusRunning = "running"
+	ExecutionStatusSuccess = "success"
+	ExecutionStatusFailed  = "failed"
+)
+
+// EventPipelineExecution 工作流执行记录
+type EventPipelineExecution struct {
+	ID           string `json:"id" gorm:"primaryKey;type:varchar(36)"`
+	PipelineID   int64  `json:"pipeline_id" gorm:"index"`
+	PipelineName string `json:"pipeline_name" gorm:"type:varchar(128)"`
+	EventID      int64  `json:"event_id" gorm:"index"`
+
+	// 触发模式：event（告警触发）、api（API触发）、cron（定时触发）
+	Mode string `json:"mode" gorm:"type:varchar(16);index"`
+
+	// 状态：running、success、failed
+	Status string `json:"status" gorm:"type:varchar(16);index"`
+
+	// 各节点执行结果（JSON）
+	NodeResults string `json:"node_results" gorm:"type:mediumtext"`
+
+	// 错误信息
+	ErrorMessage string `json:"error_message" gorm:"type:varchar(1024)"`
+	ErrorNode    string `json:"error_node" gorm:"type:varchar(36)"`
+
+	// 时间
+	CreatedAt  int64 `json:"created_at" gorm:"index"`
+	FinishedAt int64 `json:"finished_at"`
+	DurationMs int64 `json:"duration_ms"`
+
+	// 触发者信息
+	TriggerBy string `json:"trigger_by" gorm:"type:varchar(64)"`
+
+	// 环境变量快照（脱敏后存储）
+	EnvSnapshot string `json:"env_snapshot,omitempty" gorm:"type:text"`
+}
+
+func (e *EventPipelineExecution) TableName() string {
+	return "event_pipeline_execution"
+}
+
+// SetNodeResults 设置节点执行结果（序列化为 JSON）
+func (e *EventPipelineExecution) SetNodeResults(results []*NodeExecutionResult) error {
+	data, err := json.Marshal(results)
+	if err != nil {
+		return err
+	}
+	e.NodeResults = string(data)
+	return nil
+}
+
+// GetNodeResults 获取节点执行结果（反序列化）
+func (e *EventPipelineExecution) GetNodeResults() ([]*NodeExecutionResult, error) {
+	if e.NodeResults == "" {
+		return nil, nil
+	}
+	var results []*NodeExecutionResult
+	err := json.Unmarshal([]byte(e.NodeResults), &results)
+	return results, err
+}
+
+// SetEnvSnapshot 设置环境变量快照（脱敏后存储）
+func (e *EventPipelineExecution) SetEnvSnapshot(env map[string]string) error {
+	data, err := json.Marshal(env)
+	if err != nil {
+		return err
+	}
+	e.EnvSnapshot = string(data)
+	return nil
+}
+
+// GetEnvSnapshot 获取环境变量快照
+func (e *EventPipelineExecution) GetEnvSnapshot() (map[string]string, error) {
+	if e.EnvSnapshot == "" {
+		return nil, nil
+	}
+	var env map[string]string
+	err := json.Unmarshal([]byte(e.EnvSnapshot), &env)
+	return env, err
+}
+
+// CreateEventPipelineExecution 创建执行记录
+func CreateEventPipelineExecution(c *ctx.Context, execution *EventPipelineExecution) error {
+	return DB(c).Create(execution).Error
+}
+
+// UpdateEventPipelineExecution 更新执行记录
+func UpdateEventPipelineExecution(c *ctx.Context, execution *EventPipelineExecution) error {
+	return DB(c).Save(execution).Error
+}
+
+// GetEventPipelineExecution 获取单条执行记录
+func GetEventPipelineExecution(c *ctx.Context, id string) (*EventPipelineExecution, error) {
+	var execution EventPipelineExecution
+	err := DB(c).Where("id = ?", id).First(&execution).Error
+	if err != nil {
+		return nil, err
+	}
+	return &execution, nil
+}
+
+// ListEventPipelineExecutions 获取 Pipeline 的执行记录列表
+func ListEventPipelineExecutions(c *ctx.Context, pipelineID int64, mode, status string, limit, offset int) ([]*EventPipelineExecution, int64, error) {
+	var executions []*EventPipelineExecution
+	var total int64
+
+	session := DB(c).Model(&EventPipelineExecution{}).Where("pipeline_id = ?", pipelineID)
+
+	if mode != "" {
+		session = session.Where("mode = ?", mode)
+	}
+	if status != "" {
+		session = session.Where("status = ?", status)
+	}
+
+	err := session.Count(&total).Error
+	if err != nil {
+		return nil, 0, err
+	}
+
+	err = session.Order("created_at desc").Limit(limit).Offset(offset).Find(&executions).Error
+	if err != nil {
+		return nil, 0, err
+	}
+
+	return executions, total, nil
+}
+
+// ListEventPipelineExecutionsByEventID 根据事件ID获取执行记录
+func ListEventPipelineExecutionsByEventID(c *ctx.Context, eventID int64) ([]*EventPipelineExecution, error) {
+	var executions []*EventPipelineExecution
+	err := DB(c).Where("event_id = ?", eventID).Order("created_at desc").Find(&executions).Error
+	return executions, err
+}
+
+// ListAllEventPipelineExecutions 获取所有 Pipeline 的执行记录列表
+func ListAllEventPipelineExecutions(c *ctx.Context, pipelineName, mode, status string, limit, offset int) ([]*EventPipelineExecution, int64, error) {
+	var executions []*EventPipelineExecution
+	var total int64
+
+	session := DB(c).Model(&EventPipelineExecution{})
+
+	if pipelineName != "" {
+		session = session.Where("pipeline_name LIKE ?", "%"+pipelineName+"%")
+	}
+	if mode != "" {
+		session = session.Where("mode = ?", mode)
+	}
+	if status != "" {
+		session = session.Where("status = ?", status)
+	}
+
+	err := session.Count(&total).Error
+	if err != nil {
+		return nil, 0, err
+	}
+
+	err = session.Order("created_at desc").Limit(limit).Offset(offset).Find(&executions).Error
+	if err != nil {
+		return nil, 0, err
+	}
+
+	return executions, total, nil
+}
+
+// DeleteEventPipelineExecutions 批量删除执行记录（按时间）
+func DeleteEventPipelineExecutions(c *ctx.Context, beforeTime int64) (int64, error) {
+	result := DB(c).Where("created_at < ?", beforeTime).Delete(&EventPipelineExecution{})
+	return result.RowsAffected, result.Error
+}
+
+// DeleteEventPipelineExecutionsByPipelineID 删除指定 Pipeline 的所有执行记录
+func DeleteEventPipelineExecutionsByPipelineID(c *ctx.Context, pipelineID int64) error {
+	return DB(c).Where("pipeline_id = ?", pipelineID).Delete(&EventPipelineExecution{}).Error
+}
+
+// EventPipelineExecutionStatistics 执行统计
+type EventPipelineExecutionStatistics struct {
+	Total     int64 `json:"total"`
+	Success   int64 `json:"success"`
+	Failed    int64 `json:"failed"`
+	Running   int64 `json:"running"`
+	AvgDurMs  int64 `json:"avg_duration_ms"`
+	LastRunAt int64 `json:"last_run_at"`
+}
+
+// GetEventPipelineExecutionStatistics 获取执行统计信息
+func GetEventPipelineExecutionStatistics(c *ctx.Context, pipelineID int64) (*EventPipelineExecutionStatistics, error) {
+	var stats EventPipelineExecutionStatistics
+
+	// 总数
+	err := DB(c).Model(&EventPipelineExecution{}).Where("pipeline_id = ?", pipelineID).Count(&stats.Total).Error
+	if err != nil {
+		return nil, err
+	}
+
+	// 成功数
+	err = DB(c).Model(&EventPipelineExecution{}).Where("pipeline_id = ? AND status = ?", pipelineID, ExecutionStatusSuccess).Count(&stats.Success).Error
+	if err != nil {
+		return nil, err
+	}
+
+	// 失败数
+	err = DB(c).Model(&EventPipelineExecution{}).Where("pipeline_id = ? AND status = ?", pipelineID, ExecutionStatusFailed).Count(&stats.Failed).Error
+	if err != nil {
+		return nil, err
+	}
+
+	// 运行中
+	err = DB(c).Model(&EventPipelineExecution{}).Where("pipeline_id = ? AND status = ?", pipelineID, ExecutionStatusRunning).Count(&stats.Running).Error
+	if err != nil {
+		return nil, err
+	}
+
+	// 平均耗时
+	var avgDur struct {
+		AvgDur float64 `gorm:"column:avg_dur"`
+	}
+	err = DB(c).Model(&EventPipelineExecution{}).
+		Select("AVG(duration_ms) as avg_dur").
+		Where("pipeline_id = ? AND status = ?", pipelineID, ExecutionStatusSuccess).
+		Scan(&avgDur).Error
+	if err != nil {
+		return nil, err
+	}
+	stats.AvgDurMs = int64(avgDur.AvgDur)
+
+	// 最后执行时间
+	var lastExec EventPipelineExecution
+	err = DB(c).Where("pipeline_id = ?", pipelineID).Order("created_at desc").First(&lastExec).Error
+	if err == nil {
+		stats.LastRunAt = lastExec.CreatedAt
+	}
+
+	return &stats, nil
+}
+
+// EventPipelineExecutionDetail 执行详情（包含解析后的节点结果）
+type EventPipelineExecutionDetail struct {
+	EventPipelineExecution
+	NodeResultsParsed []*NodeExecutionResult `json:"node_results_parsed"`
+	EnvSnapshotParsed map[string]string      `json:"env_snapshot_parsed"`
+}
+
+// GetEventPipelineExecutionDetail 获取执行详情
+func GetEventPipelineExecutionDetail(c *ctx.Context, id string) (*EventPipelineExecutionDetail, error) {
+	execution, err := GetEventPipelineExecution(c, id)
+	if err != nil {
+		return nil, err
+	}
+
+	detail := &EventPipelineExecutionDetail{
+		EventPipelineExecution: *execution,
+	}
+
+	// 解析节点结果
+	nodeResults, err := execution.GetNodeResults()
+	if err != nil {
+		return nil, fmt.Errorf("parse node results error: %w", err)
+	}
+	detail.NodeResultsParsed = nodeResults
+
+	// 解析环境变量快照
+	envSnapshot, err := execution.GetEnvSnapshot()
+	if err != nil {
+		return nil, fmt.Errorf("parse env snapshot error: %w", err)
+	}
+	detail.EnvSnapshotParsed = envSnapshot
+
+	return detail, nil
+}

models/event_processor.go

@@ -9,11 +9,21 @@ import (
 
 type Processor interface {
 	Init(settings interface{}) (Processor, error) // 初始化配置
-	Process(ctx *ctx.Context, event *AlertCurEvent) (*AlertCurEvent, string, error)
+	Process(ctx *ctx.Context, wfCtx *WorkflowContext) (*WorkflowContext, string, error)
 	// 处理器有三种情况：
-	// 1. 处理成功，返回处理后的事件
-	// 2. 处理成功，不需要返回处理后端事件，只返回处理结果，将处理结果放到 string 中，比如 eventdrop callback 处理器
+	// 1. 处理成功，返回处理后的 WorkflowContext
+	// 2. 处理成功，不需要返回处理后的上下文，只返回处理结果，将处理结果放到 string 中，比如 eventdrop callback 处理器
 	// 3. 处理失败，返回错误，将错误放到 error 中
+	// WorkflowContext 包含：Event（事件）、Env（环境变量/输入参数）、Metadata（执行元数据）
+}
+
+// BranchProcessor 分支处理器接口
+// 用于 if、switch、foreach 等需要返回分支索引或特殊输出的处理器
+type BranchProcessor interface {
+	Processor
+	// ProcessWithBranch 处理事件并返回 NodeOutput
+	// NodeOutput 包含：处理后的上下文、消息、是否终止、分支索引
+	ProcessWithBranch(ctx *ctx.Context, wfCtx *WorkflowContext) (*NodeOutput, error)
 }
 
 type NewProcessorFn func(settings interface{}) (Processor, error)

models/message_tpl.go

@@ -622,9 +622,26 @@ var NewTplMap = map[string]string{
 {{if $event.RuleNote }}**Alarm description:** **{{$event.RuleNote}}**{{end}}   
 {{- end -}}
 [Event Details]({{.domain}}/share/alert-his-events/{{$event.Id}})|[Block for 1 hour]({{.domain}}/alert-mutes/add?__event_id={{$event.Id}})|[View Curve]({{.domain}}/metric/explorer?__event_id={{$event.Id}}&mode=graph)`,
+
+	// Jira and JSMAlert share the same template format
+	Jira: `Severity: S{{$event.Severity}} {{if $event.IsRecovered}}Recovered{{else}}Triggered{{end}}
+Rule Name: {{$event.RuleName}}{{if $event.RuleNote}}
+Rule Notes: {{$event.RuleNote}}{{end}}
+Metrics: {{$event.TagsJSON}}
+Annotations:
+{{- range $key, $val := $event.AnnotationsJSON}}
+{{$key}}: {{$val}}
+{{- end}}\n{{if $event.IsRecovered}}Recovery Time: {{timeformat $event.LastEvalTime}}{{else}}Trigger Time: {{timeformat $event.TriggerTime}}
+Trigger Value: {{$event.TriggerValue}}{{end}}
+Send Time: {{timestamp}}
+Event Details: {{.domain}}/share/alert-his-events/{{$event.Id}}
+Mute for 1 Hour: {{.domain}}/alert-mutes/add?__event_id={{$event.Id}}`,
 }
 
+// Weight 用于页面元素排序，weight 越大 排序越靠后
 var MsgTplMap = []MessageTemplate{
+	{Name: "Jira", Ident: Jira, Weight: 18, Content: map[string]string{"content": NewTplMap[Jira]}},
+	{Name: "JSMAlert", Ident: JSMAlert, Weight: 17, Content: map[string]string{"content": NewTplMap[Jira]}},
 	{Name: "Callback", Ident: "callback", Weight: 16, Content: map[string]string{"content": ""}},
 	{Name: "MattermostWebhook", Ident: MattermostWebhook, Weight: 15, Content: map[string]string{"content": NewTplMap[MattermostWebhook]}},
 	{Name: "MattermostBot", Ident: MattermostBot, Weight: 14, Content: map[string]string{"content": NewTplMap[MattermostWebhook]}},

models/migrate/migrate.go

@@ -68,7 +68,7 @@ func MigrateTables(db *gorm.DB) error {
 		&Board{}, &BoardBusigroup{}, &Users{}, &SsoConfig{}, &models.BuiltinMetric{},
 		&models.MetricFilter{}, &models.NotificationRecord{}, &models.TargetBusiGroup{},
 		&models.UserToken{}, &models.DashAnnotation{}, MessageTemplate{}, NotifyRule{}, NotifyChannelConfig{}, &EsIndexPatternMigrate{},
-		&models.EventPipeline{}, &models.EmbeddedProduct{}, &models.SourceToken{},
+		&models.EventPipeline{}, &models.EventPipelineExecution{}, &models.EmbeddedProduct{}, &models.SourceToken{},
 		&models.SavedView{}, &models.UserViewFavorite{}}
 
 	if isPostgres(db) {
@@ -99,7 +99,11 @@ func MigrateTables(db *gorm.DB) error {
 	}()
 
 	if !db.Migrator().HasTable(&models.BuiltinPayload{}) {
-		dts = append(dts, &models.BuiltinPayload{})
+		if isPostgres(db) {
+			dts = append(dts, &models.PostgresBuiltinPayload{})
+		} else {
+			dts = append(dts, &models.BuiltinPayload{})
+		}
 	} else {
 		dts = append(dts, &BuiltinPayloads{})
 	}

models/notify_channel.go

@@ -725,10 +725,10 @@ func (ncc *NotifyChannelConfig) SendHTTP(events []*AlertCurEvent, tpl map[string
 			logger.Errorf("send_http: failed to read response. url=%s request_body=%s error=%v", url, string(body), err)
 		}
 		if resp.StatusCode == http.StatusOK {
-			return string(body), nil
+			return fmt.Sprintf("status_code:%d, response:%s", resp.StatusCode, string(body)), nil
 		}
 
-		return "", fmt.Errorf("failed to send request, status code: %d, body: %s", resp.StatusCode, string(body))
+		return fmt.Sprintf("status_code:%d, response:%s", resp.StatusCode, string(body)), fmt.Errorf("failed to send request, status code: %d, body: %s", resp.StatusCode, string(body))
 	}
 
 	return lastErrorMessage, errors.New("all retries failed, last error: " + lastErrorMessage)
@@ -1204,24 +1204,56 @@ func (c NotiChList) IfUsed(nr *NotifyRule) bool {
 	return false
 }
 
+// Weight 用于页面元素排序，weight 越大 排序越靠后
 var NotiChMap = []*NotifyChannelConfig{
 	{
-		Name: "Callback", Ident: "callback", RequestType: "http", Weight: 2, Enable: true,
+		Name: "PagerDuty", Ident: "pagerduty", RequestType: "pagerduty", Weight: 19, Enable: true,
+		RequestConfig: &RequestConfig{
+			PagerDutyRequestConfig: &PagerDutyRequestConfig{
+				ApiKey:     "pagerduty api key",
+				Timeout:    5000,
+				RetryTimes: 3,
+			},
+		},
+	},
+	{
+		Name: "JIRA", Ident: Jira, RequestType: "http", Weight: 18, Enable: true,
 		RequestConfig: &RequestConfig{
 			HTTPRequestConfig: &HTTPRequestConfig{
-				URL:    "{{$params.callback_url}}",
-				Method: "POST", Headers: map[string]string{"Content-Type": "application/json"},
+				URL:     "https://{JIRA Service Account Email}:{API Token}@api.atlassian.com/ex/jira/{CloudID}/rest/api/3/issue",
+				Method:  "POST",
+				Headers: map[string]string{"Content-Type": "application/json"},
 				Timeout: 10000, Concurrency: 5, RetryTimes: 3, RetryInterval: 100,
 				Request: RequestDetail{
-					Body: `{{ jsonMarshal $events }}`,
+					Body: `{"fields":{"project":{"key":"{{$params.project_key}}"},"issuetype":{"name":"{{if $event.IsRecovered}}Recovery{{else}}Alert{{end}}"},"summary":"{{$event.RuleName}}","description":{"type":"doc","version":1,"content":[{"type":"paragraph","content":[{"type":"text","text":"{{$tpl.content}}"}]}]},"labels":["{{join $event.TagsJSON "\",\""}}", "eventHash={{$event.Hash}}"]}}`,
 				},
 			},
 		},
 		ParamConfig: &NotifyParamConfig{
 			Custom: Params{
 				Params: []ParamItem{
-					{Key: "callback_url", CName: "Callback Url", Type: "string"},
-					{Key: "note", CName: "Note", Type: "string"},
+					{Key: "project_key", CName: "Project Key", Type: "string"},
+				},
+			},
+		},
+	},
+	{
+		Name: "JSM Alert", Ident: JSMAlert, RequestType: "http", Weight: 17, Enable: true,
+		RequestConfig: &RequestConfig{
+			HTTPRequestConfig: &HTTPRequestConfig{
+				URL:     `https://api.atlassian.com/jsm/ops/integration/v2/alerts{{if $event.IsRecovered}}/{{$event.Hash}}/close?identifierType=alias{{else}}{{end}}`,
+				Method:  "POST",
+				Headers: map[string]string{"Content-Type": "application/json", "Authorization": "GenieKey {{$params.api_key}}"},
+				Timeout: 10000, Concurrency: 5, RetryTimes: 3, RetryInterval: 100,
+				Request: RequestDetail{
+					Body: `{{if $event.IsRecovered}}{"note":"{{$tpl.content}}","source":"{{$event.Cluster}}"}{{else}}{"message":"{{$event.RuleName}}","description":"{{$tpl.content}}","alias":"{{$event.Hash}}","priority":"P{{$event.Severity}}","tags":[{{range $i, $v := $event.TagsJSON}}{{if $i}},{{end}}"{{$v}}"{{end}}],"details":{{jsonMarshal $event.AnnotationsJSON}},"entity":"{{$event.TargetIdent}}","source":"{{$event.Cluster}}"}{{end}}`,
+				},
+			},
+		},
+		ParamConfig: &NotifyParamConfig{
+			Custom: Params{
+				Params: []ParamItem{
+					{Key: "api_key", CName: "API Key", Type: "string"},
 				},
 			},
 		},
@@ -1614,6 +1646,27 @@ var NotiChMap = []*NotifyChannelConfig{
 			},
 		},
 	},
+	{
+		Name: "Callback", Ident: "callback", RequestType: "http", Weight: 2, Enable: true,
+		RequestConfig: &RequestConfig{
+			HTTPRequestConfig: &HTTPRequestConfig{
+				URL:    "{{$params.callback_url}}",
+				Method: "POST", Headers: map[string]string{"Content-Type": "application/json"},
+				Timeout: 10000, Concurrency: 5, RetryTimes: 3, RetryInterval: 100,
+				Request: RequestDetail{
+					Body: `{{ jsonMarshal $events }}`,
+				},
+			},
+		},
+		ParamConfig: &NotifyParamConfig{
+			Custom: Params{
+				Params: []ParamItem{
+					{Key: "callback_url", CName: "Callback Url", Type: "string"},
+					{Key: "note", CName: "Note", Type: "string"},
+				},
+			},
+		},
+	},
 	{
 		Name: "FlashDuty", Ident: "flashduty", RequestType: "flashduty", Weight: 1, Enable: true,
 		RequestConfig: &RequestConfig{
@@ -1630,16 +1683,6 @@ var NotiChMap = []*NotifyChannelConfig{
 			},
 		},
 	},
-	{
-		Name: "PagerDuty", Ident: "pagerduty", RequestType: "pagerduty", Weight: 1, Enable: true,
-		RequestConfig: &RequestConfig{
-			PagerDutyRequestConfig: &PagerDutyRequestConfig{
-				ApiKey:     "pagerduty api key",
-				Timeout:    5000,
-				RetryTimes: 3,
-			},
-		},
-	},
 }
 
 func InitNotifyChannel(ctx *ctx.Context) {

models/ts.go

@@ -68,7 +68,7 @@ type RelationKey struct {
 type QueryParam struct {
 	Cate         string        `json:"cate"`
 	DatasourceId int64         `json:"datasource_id"`
-	Queries       []interface{} `json:"query"`
+	Queries      []interface{} `json:"query"`
 }
 
 type Series struct {

models/user.go

@@ -42,6 +42,8 @@ const (
 	Lark              = "lark"
 	LarkCard          = "larkcard"
 	Phone             = "phone"
+	Jira              = "jira"
+	JSMAlert          = "jsm_alert"
 
 	DingtalkKey  = "dingtalk_robot_token"
 	WecomKey     = "wecom_robot_token"
@@ -339,6 +341,11 @@ func (u *User) Del(ctx *ctx.Context) error {
 }
 
 func (u *User) ChangePassword(ctx *ctx.Context, oldpass, newpass string) error {
+	// SSO 用户（ldap/oidc/cas/oauth2/dingtalk等）且未设置本地密码，不支持本地修改密码
+	if u.Belong != "" && u.Password == "******" {
+		return fmt.Errorf("SSO user(%s) cannot change password locally, please change password in %s", u.Username, u.Belong)
+	}
+
 	_oldpass, err := CryptoPass(ctx, oldpass)
 	if err != nil {
 		return err

models/workflow.go

@@ -0,0 +1,160 @@
+package models
+
+// WorkflowNode 工作流节点
+type WorkflowNode struct {
+	ID       string      `json:"id"`                 // 节点唯一ID
+	Name     string      `json:"name"`               // 显示名称
+	Type     string      `json:"type"`               // 节点类型（对应 Processor typ）
+	Position []float64   `json:"position,omitempty"` // [x, y] UI位置
+	Config   interface{} `json:"config"`             // 节点配置
+
+	// 执行控制
+	Disabled       bool `json:"disabled,omitempty"`
+	ContinueOnFail bool `json:"continue_on_fail,omitempty"`
+	RetryOnFail    bool `json:"retry_on_fail,omitempty"`
+	MaxRetries     int  `json:"max_retries,omitempty"`
+	RetryInterval  int  `json:"retry_interval,omitempty"` // 秒
+}
+
+// Connections 节点连接关系 map[源节点ID]NodeConnections
+type Connections map[string]NodeConnections
+
+// NodeConnections 单个节点的输出连接
+type NodeConnections struct {
+	// Main 输出端口的连接
+	// Main[outputIndex] = []ConnectionTarget
+	Main [][]ConnectionTarget `json:"main"`
+}
+
+// ConnectionTarget 连接目标
+type ConnectionTarget struct {
+	Node  string `json:"node"`  // 目标节点ID
+	Type  string `json:"type"`  // 输入类型，通常是 "main"
+	Index int    `json:"index"` // 目标节点的输入端口索引
+}
+
+// EnvVariable 环境变量
+type EnvVariable struct {
+	Key         string `json:"key"`                   // 变量名
+	Value       string `json:"value"`                 // 默认值
+	Description string `json:"description,omitempty"` // 描述
+	Secret      bool   `json:"secret,omitempty"`      // 是否敏感（日志脱敏）
+	Required    bool   `json:"required,omitempty"`    // 是否必填
+}
+
+// NodeOutput 节点执行输出
+type NodeOutput struct {
+	WfCtx       *WorkflowContext `json:"wf_ctx"`                 // 处理后的工作流上下文
+	Message     string           `json:"message"`                // 处理消息
+	Terminate   bool             `json:"terminate"`              // 是否终止流程
+	BranchIndex *int             `json:"branch_index,omitempty"` // 分支索引（条件节点使用）
+
+	// 流式输出支持
+	Stream     bool              `json:"stream,omitempty"` // 是否流式输出
+	StreamChan chan *StreamChunk `json:"-"`                // 流式数据通道（不序列化）
+}
+
+// WorkflowResult 工作流执行结果
+type WorkflowResult struct {
+	Event       *AlertCurEvent         `json:"event"`        // 最终事件
+	Status      string                 `json:"status"`       // success, failed, streaming
+	Message     string                 `json:"message"`      // 汇总消息
+	NodeResults []*NodeExecutionResult `json:"node_results"` // 各节点执行结果
+	ErrorNode   string                 `json:"error_node,omitempty"`
+
+	// 流式输出支持
+	Stream     bool              `json:"stream,omitempty"` // 是否流式输出
+	StreamChan chan *StreamChunk `json:"-"`                // 流式数据通道（不序列化）
+}
+
+// NodeExecutionResult 节点执行结果
+type NodeExecutionResult struct {
+	NodeID      string `json:"node_id"`
+	NodeName    string `json:"node_name"`
+	NodeType    string `json:"node_type"`
+	Status      string `json:"status"` // success, failed, skipped
+	Message     string `json:"message"`
+	StartedAt   int64  `json:"started_at"`
+	FinishedAt  int64  `json:"finished_at"`
+	DurationMs  int64  `json:"duration_ms"`
+	Error       string `json:"error,omitempty"`
+	BranchIndex *int   `json:"branch_index,omitempty"` // 条件节点的分支选择
+}
+
+// 触发模式常量
+const (
+	TriggerModeEvent = "event" // 告警事件触发
+	TriggerModeAPI   = "api"   // API 触发
+	TriggerModeCron  = "cron"  // 定时触发（后续支持）
+)
+
+// WorkflowTriggerContext 工作流触发上下文
+type WorkflowTriggerContext struct {
+	// 触发模式
+	Mode string `json:"mode"`
+
+	// 触发者
+	TriggerBy string `json:"trigger_by"`
+
+	// 请求ID（API/Cron 触发使用）
+	RequestID string `json:"request_id"`
+
+	// 环境变量覆盖
+	EnvOverrides map[string]string `json:"env_overrides"`
+
+	// 流式输出（API 调用时动态指定）
+	Stream bool `json:"stream"`
+
+	// Cron 相关（后续使用）
+	CronJobID   string `json:"cron_job_id,omitempty"`
+	CronExpr    string `json:"cron_expr,omitempty"`
+	ScheduledAt int64  `json:"scheduled_at,omitempty"`
+}
+
+type WorkflowContext struct {
+	Event    *AlertCurEvent         `json:"event"`            // 当前事件
+	Env      map[string]string      `json:"env"`              // 环境变量/配置（静态，来自 Pipeline 配置）
+	Vars     map[string]interface{} `json:"vars"`             // 节点间传递的数据（动态，运行时产生）
+	Metadata map[string]string      `json:"metadata"`         // 执行元数据（request_id、start_time 等）
+	Output   map[string]interface{} `json:"output,omitempty"` // 输出结果（非告警场景使用）
+
+	// 流式输出支持
+	Stream     bool              `json:"-"` // 是否启用流式输出（不序列化）
+	StreamChan chan *StreamChunk `json:"-"` // 流式数据通道（不序列化）
+}
+
+// SanitizedEnv 返回脱敏后的环境变量（用于日志和存储）
+func (ctx *WorkflowContext) SanitizedEnv(secretKeys map[string]bool) map[string]string {
+	sanitized := make(map[string]string)
+	for k, v := range ctx.Env {
+		if secretKeys[k] {
+			sanitized[k] = "******"
+		} else {
+			sanitized[k] = v
+		}
+	}
+	return sanitized
+}
+
+// StreamChunk 类型常量
+const (
+	StreamTypeThinking   = "thinking"    // AI 思考过程（ReAct Thought）
+	StreamTypeToolCall   = "tool_call"   // 工具调用
+	StreamTypeToolResult = "tool_result" // 工具执行结果
+	StreamTypeText       = "text"        // LLM 文本输出
+	StreamTypeDone       = "done"        // 完成
+	StreamTypeError      = "error"       // 错误
+)
+
+// StreamChunk 流式数据块
+type StreamChunk struct {
+	Type      string      `json:"type"`                 // thinking / tool_call / tool_result / text / done / error
+	Content   string      `json:"content"`              // 完整内容（累积）
+	Delta     string      `json:"delta,omitempty"`      // 增量内容
+	NodeID    string      `json:"node_id,omitempty"`    // 当前节点 ID
+	RequestID string      `json:"request_id,omitempty"` // 请求追踪 ID
+	Metadata  interface{} `json:"metadata,omitempty"`   // 额外元数据（如工具调用参数）
+	Done      bool        `json:"done"`                 // 是否结束
+	Error     string      `json:"error,omitempty"`      // 错误信息
+	Timestamp int64       `json:"timestamp"`            // 时间戳(毫秒)
+}

pkg/ormx/database_init.go

@@ -1084,7 +1084,7 @@ type InitPostgresDatasource struct {
 	Status         string `gorm:"size:255;not null;default:''"`
 	HTTP           string `gorm:"size:4096;not null;default:''"`
 	Auth           string `gorm:"size:8192;not null;default:''"`
-	IsDefault      bool   `gorm:"typr:boolean;not null;default:0"`
+	IsDefault      bool   `gorm:"type:boolean;not null;default:0"`
 	CreatedAt      int64  `gorm:"not null;default:0"`
 	CreatedBy      string `gorm:"size:64;not null;default:''"`
 	UpdatedAt      int64  `gorm:"not null;default:0"`
@@ -1494,10 +1494,6 @@ func sqliteDataBaseInit(db *gorm.DB) error {
 		{RoleName: "Standard", Operation: "/alert-rules-built-in"},
 		{RoleName: "Standard", Operation: "/dashboards-built-in"},
 		{RoleName: "Standard", Operation: "/trace/dependencies"},
-		{RoleName: "Admin", Operation: "/help/source"},
-		{RoleName: "Admin", Operation: "/help/sso"},
-		{RoleName: "Admin", Operation: "/help/notification-tpls"},
-		{RoleName: "Admin", Operation: "/help/notification-settings"},
 		{RoleName: "Standard", Operation: "/users"},
 		{RoleName: "Standard", Operation: "/user-groups"},
 		{RoleName: "Standard", Operation: "/user-groups/add"},
@@ -1659,8 +1655,7 @@ func mysqlDataBaseInit(db *gorm.DB) error {
 	for _, dt := range dts {
 		err := db.AutoMigrate(dt)
 		if err != nil {
-			fmt.Printf("mysqlDataBaseInit AutoMigrate error: %v\n", err)
-			return err
+			logger.Errorf("mysqlDataBaseInit AutoMigrate error: %v\n", err)
 		}
 	}
 
@@ -1668,7 +1663,7 @@ func mysqlDataBaseInit(db *gorm.DB) error {
 		tableName := "task_host_" + strconv.Itoa(i)
 		err := db.Table(tableName).AutoMigrate(&InitTaskHost{})
 		if err != nil {
-			return err
+			logger.Errorf("mysqlDataBaseInit AutoMigrate task_host_%d error: %v\n", i, err)
 		}
 	}
 
@@ -1690,10 +1685,6 @@ func mysqlDataBaseInit(db *gorm.DB) error {
 		{RoleName: "Standard", Operation: "/alert-rules-built-in"},
 		{RoleName: "Standard", Operation: "/dashboards-built-in"},
 		{RoleName: "Standard", Operation: "/trace/dependencies"},
-		{RoleName: "Admin", Operation: "/help/source"},
-		{RoleName: "Admin", Operation: "/help/sso"},
-		{RoleName: "Admin", Operation: "/help/notification-tpls"},
-		{RoleName: "Admin", Operation: "/help/notification-settings"},
 		{RoleName: "Standard", Operation: "/users"},
 		{RoleName: "Standard", Operation: "/user-groups"},
 		{RoleName: "Standard", Operation: "/user-groups/add"},
@@ -1886,10 +1877,6 @@ func postgresDataBaseInit(db *gorm.DB) error {
 		{RoleName: "Standard", Operation: "/alert-rules-built-in"},
 		{RoleName: "Standard", Operation: "/dashboards-built-in"},
 		{RoleName: "Standard", Operation: "/trace/dependencies"},
-		{RoleName: "Admin", Operation: "/help/source"},
-		{RoleName: "Admin", Operation: "/help/sso"},
-		{RoleName: "Admin", Operation: "/help/notification-tpls"},
-		{RoleName: "Admin", Operation: "/help/notification-settings"},
 		{RoleName: "Standard", Operation: "/users"},
 		{RoleName: "Standard", Operation: "/user-groups"},
 		{RoleName: "Standard", Operation: "/user-groups/add"},

pkg/ormx/ormx.go

@@ -348,6 +348,15 @@ func New(c DBConfig) (*gorm.DB, error) {
 		return nil, fmt.Errorf("failed to open database: %v", err)
 	}
 
+	// 检查 user 表是否存在，可能用户自己创建了空的数据库，如果不存在也执行 DataBaseInit
+	if dbExist && !db.Migrator().HasTable("users") {
+		fmt.Printf("Database exists but user table not found, initializing tables for %s\n", c.DBType)
+		err = DataBaseInit(c, db)
+		if err != nil {
+			return nil, fmt.Errorf("failed to init database: %v", err)
+		}
+	}
+
 	if c.Debug {
 		db = db.Debug()
 	}

prom/option.go

@@ -3,7 +3,7 @@ package prom
 import (
 	"sync"
 
-	"github.com/ccfos/nightingale/v6/pkg/tlsx"
+	"github.com/ccfos/nightingale/v6/models"
 )
 
 type PromOption struct {
@@ -20,7 +20,8 @@ type PromOption struct {
 
 	Headers []string
 
-	tlsx.ClientConfig
+	// TLS 配置（支持 mTLS）
+	TLS models.TLS
 }
 
 func (po *PromOption) Equal(target PromOption) bool {
@@ -52,10 +53,6 @@ func (po *PromOption) Equal(target PromOption) bool {
 		return false
 	}
 
-	if po.InsecureSkipVerify != target.InsecureSkipVerify {
-		return false
-	}
-
 	if len(po.Headers) != len(target.Headers) {
 		return false
 	}
@@ -66,6 +63,29 @@ func (po *PromOption) Equal(target PromOption) bool {
 		}
 	}
 
+	// 比较 TLS 配置
+	if po.TLS.SkipTlsVerify != target.TLS.SkipTlsVerify {
+		return false
+	}
+	if po.TLS.CACert != target.TLS.CACert {
+		return false
+	}
+	if po.TLS.ClientCert != target.TLS.ClientCert {
+		return false
+	}
+	if po.TLS.ClientKey != target.TLS.ClientKey {
+		return false
+	}
+	if po.TLS.ServerName != target.TLS.ServerName {
+		return false
+	}
+	if po.TLS.MinVersion != target.TLS.MinVersion {
+		return false
+	}
+	if po.TLS.MaxVersion != target.TLS.MaxVersion {
+		return false
+	}
+
 	return true
 }
 

prom/reader.go

@@ -101,11 +101,7 @@ func (pc *PromClientMap) loadFromDatabase() {
 			DialTimeout:         ds.HTTPJson.DialTimeout,
 			MaxIdleConnsPerHost: ds.HTTPJson.MaxIdleConnsPerHost,
 			Headers:             header,
-		}
-
-		if strings.HasPrefix(ds.HTTPJson.Url, "https") {
-			po.UseTLS = true
-			po.InsecureSkipVerify = ds.HTTPJson.TLS.SkipTlsVerify
+			TLS:                 ds.HTTPJson.TLS,
 		}
 
 		if internalAddr != "" && !pc.ctx.IsCenter {
@@ -149,7 +145,10 @@ func (pc *PromClientMap) loadFromDatabase() {
 }
 
 func (pc *PromClientMap) newReaderClientFromPromOption(po PromOption) (api.Client, error) {
-	tlsConfig, _ := po.TLSConfig()
+	tlsConfig, err := po.TLS.TLSConfig()
+	if err != nil {
+		return nil, fmt.Errorf("failed to create TLS config: %v", err)
+	}
 
 	return api.NewClient(api.Config{
 		Address: po.Url,
@@ -166,7 +165,10 @@ func (pc *PromClientMap) newReaderClientFromPromOption(po PromOption) (api.Clien
 }
 
 func (pc *PromClientMap) newWriterClientFromPromOption(po PromOption) (api.Client, error) {
-	tlsConfig, _ := po.TLSConfig()
+	tlsConfig, err := po.TLS.TLSConfig()
+	if err != nil {
+		return nil, fmt.Errorf("failed to create TLS config: %v", err)
+	}
 
 	return api.NewClient(api.Config{
 		Address: po.WriteAddr,

pushgw/router/router.go

@@ -32,7 +32,7 @@ type Router struct {
 	Writers        *writer.WritersType
 	Ctx            *ctx.Context
 	HandleTS       HandleTSFunc
-	HeartbeatApi  string
+	HeartbeatApi   string
 }
 
 func stat() gin.HandlerFunc {