feat: add traceable scoped AIOps diagnosis

src/main/java/com/superbiz/agent/controller/ChatController.java

@@ -4,6 +4,7 @@ import com.alibaba.cloud.ai.graph.OverAllState;
 import lombok.Getter;
 import lombok.Setter;
 import com.superbiz.agent.domain.model.SessionContext;
+import com.superbiz.agent.dto.AIOpsRequest;
 import com.superbiz.agent.service.AiOpsService;
 import com.superbiz.agent.service.ChatService;
 import com.superbiz.agent.service.session.SessionManager;
@@ -212,21 +213,23 @@ public class ChatController {
      * 无需用户输入，自动执行告警分析流程
      */
     @PostMapping(value = "/ai_ops", produces = "text/event-stream;charset=UTF-8")
-    public SseEmitter aiOps() {
+    public SseEmitter aiOps(@RequestBody(required = false) AIOpsRequest request) {
         SseEmitter emitter = new SseEmitter(600000L); // 10分钟超时（告警分析可能较慢）
+        String sessionId = aiOpsService.resolveSessionId(request);
 
         executor.execute(() -> {
             try {
-                logger.info("收到 AI 智能运维请求 - 启动多 Agent 协作流程");
+                logger.info("收到 AI 智能运维请求 - SessionId: {}, 启动多 Agent 协作流程", sessionId);
 
                 ChatModel chatModel = chatService.getChatModel();
 
                 ToolCallback[] toolCallbacks = tools != null ? tools.getToolCallbacks() : new ToolCallback[0];
 
+                emitter.send(SseEmitter.event().name("message").data(SseMessage.session(sessionId), MediaType.APPLICATION_JSON));
                 emitter.send(SseEmitter.event().name("message").data(SseMessage.content("正在读取告警并拆解任务...\n")));
                 
                 // 调用 AiOpsService 执行分析流程
-                Optional<OverAllState> overAllStateOptional = aiOpsService.executeAiOpsAnalysis(chatModel, toolCallbacks);
+                Optional<OverAllState> overAllStateOptional = aiOpsService.executeAiOpsAnalysis(chatModel, toolCallbacks, request, sessionId);
 
                 if (overAllStateOptional.isEmpty()) {
                     emitter.send(SseEmitter.event().name("message")
@@ -245,6 +248,7 @@ public class ChatController {
                 if (finalReportOptional.isPresent()) {
                     String finalReportText = finalReportOptional.get();
                     logger.info("提取到 Planner 最终报告，长度: {}", finalReportText.length());
+                    aiOpsService.persistFinalReport(sessionId, finalReportText);
                     
                     // 发送分隔线
                     emitter.send(SseEmitter.event().name("message")
@@ -443,6 +447,13 @@ public class ChatController {
             return message;
         }
 
+        public static SseMessage session(String sessionId) {
+            SseMessage message = new SseMessage();
+            message.setType("session");
+            message.setData(sessionId);
+            return message;
+        }
+
         public static SseMessage error(String errorMessage) {
             SseMessage message = new SseMessage();
             message.setType("error");

src/main/java/com/superbiz/agent/dto/AIOpsRequest.java

@@ -7,6 +7,36 @@ import lombok.Data;
  */
 @Data
 public class AIOpsRequest {
+
+    /**
+     * 诊断会话 ID；为空时后端自动生成。
+     */
+    private String sessionId;
+
+    /**
+     * 告警名称。
+     */
+    private String alertName;
+
+    /**
+     * 受影响服务。
+     */
+    private String service;
+
+    /**
+     * 告警等级，例如 P0/P1/P2。
+     */
+    private String severity;
+
+    /**
+     * 告警描述。
+     */
+    private String description;
+
+    /**
+     * 排查时间范围，例如 last_15m。
+     */
+    private String timeRange;
     
     /**
      * 用户请求描述

src/main/java/com/superbiz/agent/repository/ToolInvocationRepository.java

@@ -22,6 +22,11 @@ public interface ToolInvocationRepository extends JpaRepository<ToolInvocation,
      */
     List<ToolInvocation> findBySessionIdOrderByIdAsc(String sessionId);
 
+    /**
+     * 根据会话ID统计真实工具调用次数
+     */
+    long countBySessionId(String sessionId);
+
     /**
      * 根据工具名查询所有调用
      */

src/main/java/com/superbiz/agent/service/AiOpsService.java

@@ -10,11 +10,12 @@ import com.superbiz.agent.agent.tool.InternalDocsTools;
 import com.superbiz.agent.agent.tool.QueryLogsTools;
 import com.superbiz.agent.agent.tool.QueryMetricsTools;
 import com.superbiz.agent.domain.entity.AgentStep;
-import com.superbiz.agent.domain.entity.AgentStep;
 import com.superbiz.agent.domain.entity.DiagnosisSession;
+import com.superbiz.agent.dto.AIOpsRequest;
 import com.superbiz.agent.hook.AgentLoggingHook;
 import com.superbiz.agent.repository.AgentStepRepository;
 import com.superbiz.agent.repository.DiagnosisSessionRepository;
+import com.superbiz.agent.repository.ToolInvocationRepository;
 import com.superbiz.agent.util.SessionContextHolder;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
@@ -62,6 +63,9 @@ public class AiOpsService {
     @Autowired
     private AgentStepRepository agentStepRepository;
 
+    @Autowired
+    private ToolInvocationRepository toolInvocationRepository;
+
     /**
      * 执行 AI Ops 告警分析流程
      *
@@ -71,22 +75,22 @@ public class AiOpsService {
      * @throws GraphRunnerException 如果 Agent 执行失败
      */
     public Optional<OverAllState> executeAiOpsAnalysis(ChatModel chatModel, ToolCallback[] toolCallbacks) throws GraphRunnerException {
+        return executeAiOpsAnalysis(chatModel, toolCallbacks, null, resolveSessionId(null));
+    }
+
+    public Optional<OverAllState> executeAiOpsAnalysis(ChatModel chatModel, ToolCallback[] toolCallbacks,
+                                                       AIOpsRequest request, String sessionId) throws GraphRunnerException {
         logger.info("开始执行 AI Ops 多 Agent 协作流程");
 
-        String sessionId = UUID.randomUUID().toString().substring(0, 8);
+        String resolvedSessionId = isBlank(sessionId) ? resolveSessionId(request) : sessionId.trim();
         long startTime = System.currentTimeMillis();
 
-        // 创建诊断会话
-        DiagnosisSession session = DiagnosisSession.builder()
-                .sessionId(sessionId)
-                .query("AI Ops 告警分析")
-                .status("RUNNING")
-                .agentFlow("AI_OPS")
-                .build();
+        // 创建或更新诊断会话
+        DiagnosisSession session = startDiagnosisSession(resolvedSessionId, request);
         diagnosisSessionRepository.save(session);
 
         // 设置 ThreadLocal 上下文（LookupKnowledgeTool 通过此获取 sessionId）
-        SessionContextHolder.setSessionId(sessionId);
+        SessionContextHolder.setSessionId(resolvedSessionId);
 
         try {
             // 构建 Planner 和 Executor Agent（每个 Agent 各自带 Hook）
@@ -102,7 +106,7 @@ public class AiOpsService {
                     .subAgents(List.of(plannerAgent, executorAgent))
                     .build();
 
-            String taskPrompt = "你是企业级 SRE，接到了自动化告警排查任务。请结合工具调用，执行**规划→执行→再规划**的闭环，并最终按照固定模板输出《告警分析报告》。禁止编造虚假数据，如连续多次查询失败需诚实反馈无法完成的原因。";
+            String taskPrompt = buildTaskPrompt(request);
 
             logger.info("调用 Supervisor Agent 开始编排...");
 
@@ -158,6 +162,87 @@ public class AiOpsService {
         }
     }
 
+    public String resolveSessionId(AIOpsRequest request) {
+        if (request != null && !isBlank(request.getSessionId())) {
+            return request.getSessionId().trim();
+        }
+        return UUID.randomUUID().toString();
+    }
+
+    public void persistFinalReport(String sessionId, String finalReport) {
+        if (isBlank(sessionId) || isBlank(finalReport)) {
+            return;
+        }
+        diagnosisSessionRepository.findBySessionId(sessionId.trim()).ifPresent(session -> {
+            session.setAnswer(finalReport);
+            diagnosisSessionRepository.save(session);
+        });
+    }
+
+    String buildQuerySummary(AIOpsRequest request) {
+        if (request == null) {
+            return "AI Ops 告警分析";
+        }
+
+        StringBuilder summary = new StringBuilder("AI Ops 告警分析");
+        appendField(summary, "告警", request.getAlertName());
+        appendField(summary, "服务", request.getService());
+        appendField(summary, "等级", request.getSeverity());
+        appendField(summary, "时间范围", request.getTimeRange());
+        appendField(summary, "描述", request.getDescription());
+        appendField(summary, "请求", request.getUserRequest());
+        return summary.toString();
+    }
+
+    boolean hasAlertPayload(AIOpsRequest request) {
+        if (request == null) {
+            return false;
+        }
+        return !isBlank(request.getAlertName())
+                || !isBlank(request.getService())
+                || !isBlank(request.getSeverity())
+                || !isBlank(request.getDescription())
+                || !isBlank(request.getTimeRange());
+    }
+
+    String buildTaskPrompt(AIOpsRequest request) {
+        StringBuilder prompt = new StringBuilder();
+        prompt.append("你是企业级 SRE，接到了自动化告警排查任务。请结合工具调用，执行**规划→执行→再规划**的闭环，并最终按照固定模板输出《告警分析报告》。禁止编造虚假数据，如连续多次查询失败需诚实反馈无法完成的原因。");
+        prompt.append("\n\n本次告警输入：\n");
+        prompt.append(buildQuerySummary(request));
+        if (hasAlertPayload(request)) {
+            prompt.append("\n\nAIOps scope mode: PAYLOAD_TARGETED\n");
+            prompt.append("- The request includes an alert payload. Treat the supplied alert payload as the primary and only main diagnosis target.\n");
+            prompt.append("- The final report must focus on the supplied alert fields such as alertName, service, severity, description, and timeRange.\n");
+            prompt.append("- You may call queryPrometheusAlerts only to verify whether the supplied alert is still active or to identify related risk/context.\n");
+            prompt.append("- If queryPrometheusAlerts returns unrelated active alerts, do not create full root-cause or remediation sections for them.\n");
+            prompt.append("- Mention unrelated active alerts only briefly in a Related Risk section when they help explain the supplied alert.\n");
+        } else {
+            prompt.append("\n\nAIOps scope mode: AUTO_DISCOVERY\n");
+            prompt.append("- The request does not include alert payload fields. First call queryPrometheusAlerts to discover current active/firing alerts.\n");
+            prompt.append("- Prefer P0/P1 alerts or the longest-running firing alerts, then diagnose one or more alerts based on severity and evidence.\n");
+            prompt.append("- Use metrics, logs, and knowledge-base evidence before producing the final alert analysis report.\n");
+        }
+        return prompt.toString();
+    }
+
+    private DiagnosisSession startDiagnosisSession(String sessionId, AIOpsRequest request) {
+        DiagnosisSession session = diagnosisSessionRepository.findBySessionId(sessionId)
+                .orElseGet(() -> DiagnosisSession.builder()
+                        .sessionId(sessionId)
+                        .agentFlow("AI_OPS")
+                        .build());
+        session.setQuery(buildQuerySummary(request));
+        session.setStatus("RUNNING");
+        session.setAgentFlow("AI_OPS");
+        session.setAnswer(null);
+        session.setTotalDurationMs(null);
+        session.setTotalTokenCount(null);
+        session.setStepCount(null);
+        session.setToolCallCount(null);
+        return session;
+    }
+
     /**
      * 构建 Planner Agent
      */
@@ -205,25 +290,33 @@ public class AiOpsService {
         }
     }
 
-    /** 从 agent_step 汇总指标回填 diagnosis_session */
+    /** 从 agent_step 和 tool_invocation 汇总指标回填 diagnosis_session */
     private void backfillSessionMetrics(DiagnosisSession session) {
         try {
             List<AgentStep> steps = agentStepRepository.findBySessionIdOrderByStepIndex(session.getSessionId());
-            if (steps.isEmpty()) return;
 
             int totalTokens = 0;
             int stepCount = 0;
-            int toolCallCount = 0;
             for (AgentStep s : steps) {
                 stepCount++;
                 if (s.getTokenCount() != null) totalTokens += s.getTokenCount();
-                if (Boolean.TRUE.equals(s.getHasToolCall())) toolCallCount++;
             }
+            long toolCallCount = toolInvocationRepository.countBySessionId(session.getSessionId());
             session.setTotalTokenCount(totalTokens);
             session.setStepCount(stepCount);
-            session.setToolCallCount(toolCallCount);
+            session.setToolCallCount(Math.toIntExact(toolCallCount));
         } catch (Exception e) {
             logger.warn("回填会话指标失败: sessionId={}", session.getSessionId(), e);
         }
     }
+
+    private void appendField(StringBuilder builder, String label, String value) {
+        if (!isBlank(value)) {
+            builder.append("\n- ").append(label).append(": ").append(value.trim());
+        }
+    }
+
+    private boolean isBlank(String value) {
+        return value == null || value.trim().isEmpty();
+    }
 }

src/main/java/com/superbiz/agent/service/ChatService.java

@@ -18,6 +18,7 @@ import com.superbiz.agent.hook.TokenUsageHolder;
 import com.superbiz.agent.hook.VerifierInputHook;
 import com.superbiz.agent.repository.AgentStepRepository;
 import com.superbiz.agent.repository.DiagnosisSessionRepository;
+import com.superbiz.agent.repository.ToolInvocationRepository;
 import com.superbiz.agent.tool.LookupKnowledgeTool;
 import com.superbiz.agent.tool.RetrievedDocTracker;
 import com.superbiz.agent.util.QuestionComplexity;
@@ -86,6 +87,9 @@ public class ChatService {
     @Autowired
     private AgentStepRepository agentStepRepository;
 
+    @Autowired
+    private ToolInvocationRepository toolInvocationRepository;
+
     @Autowired
     private EvaluationService evaluationService;
 
@@ -838,25 +842,22 @@ public class ChatService {
     ) {
     }
 
-    /** 从 agent_step 汇总 token、步数等指标回填 diagnosis_session */
+    /** 从 agent_step 和 tool_invocation 汇总指标回填 diagnosis_session */
     private void backfillSessionMetrics(DiagnosisSession session) {
         try {
             List<com.superbiz.agent.domain.entity.AgentStep> steps =
                 agentStepRepository.findBySessionIdOrderByStepIndex(session.getSessionId());
 
-            if (steps.isEmpty()) return;
-
             int totalTokens = 0;
             int stepCount = 0;
-            int toolCallCount = 0;
             for (var s : steps) {
                 stepCount++;
                 if (s.getTokenCount() != null) totalTokens += s.getTokenCount();
-                if (Boolean.TRUE.equals(s.getHasToolCall())) toolCallCount++;
             }
+            long toolCallCount = toolInvocationRepository.countBySessionId(session.getSessionId());
             session.setTotalTokenCount(totalTokens);
             session.setStepCount(stepCount);
-            session.setToolCallCount(toolCallCount);
+            session.setToolCallCount(Math.toIntExact(toolCallCount));
         } catch (Exception e) {
             logger.warn("回填会话指标失败: sessionId={}", session.getSessionId(), e);
         }

src/main/resources/application.yml

@@ -47,6 +47,14 @@ spring:
     username: root
     password: '!Fucker123..'
     driver-class-name: com.mysql.cj.jdbc.Driver
+    hikari:
+      maximum-pool-size: 5
+      minimum-idle: 1
+      connection-timeout: 10000
+      validation-timeout: 5000
+      idle-timeout: 60000
+      max-lifetime: 120000
+      keepalive-time: 30000
 
   # =====================================================
   # JPA 配置