PagerDuty 联动 Port AI 的 n8n 事件全生命周期自动化

PagerDuty webhook 事件 → Port 软件目录上下文查询 → OpenAI 严重度评估 → Slack 分级通知与 Jira 工单 → 解决后 MTTR 计算与事后分析自动化

给谁

拥有 PagerDuty、Port、Jira、Slack 账户的 SRE 或平台工程师，使用自托管 n8n 实例

你会得到

事件触发时自动拉取服务上下文、评估严重度、创建工单并通知团队；事件解决后自动计算 MTTR、生成事后分析模板、调度复盘会议并回写可靠性指标

你需要

n8n · slack · jira · port

步骤

1. 配置 PagerDuty Webhook 与事件路由

   在 n8n 中创建 Webhook 节点，路径设为 pagerduty/incident，方法为 POST。在 PagerDuty 中配置 webhook 指向该 URL。添加 Switch 节点，按 body.event.event_type 路由：incident.triggered 走触发流，incident.resolved 走解决流。

2. 查询 Port 软件目录获取事件上下文

   添加 Port 自定义节点（agentIdentifier: context_retriever_agent），输入事件 ID、标题、服务名、紧急程度和创建时间。节点向 Port 目录查询：服务层级、SLA、负责人、值班工程师、近 48 小时部署、依赖关系、Runbook 链接、监控面板、历史事件。输出为指定 JSON 格式，缺失字段留空。

   运行提示词 · 来源原文。Port 自定义节点 'Extract Context (New Incident)' 的 prompt 字段

   复制提示词

   =A PagerDuty incident has been triggered.
   
   Incident Details:
   - ID: {{ $('On PagerDuty Incident').item.json.body.event.data.id }}
   - Title: {{ $('On PagerDuty Incident').item.json.body.event.data.title }}
   - Service: {{ $('On PagerDuty Incident').item.json.body.event.data.service.name }}
   - Urgency: {{ $('On PagerDuty Incident').item.json.body.event.data.urgency }}
   - Created At: {{ $('On PagerDuty Incident').item.json.body.event.data.created_at }}
   
   Query the Port catalog for comprehensive context about this incident.
   
   Retrieve:
   1. Affected service details (tier, criticality, SLA)
   2. Service owners and current on-call engineers
   3. Recent deployments (last 48 hours)
   4. Related pull requests or configuration changes
   5. Dependencies and upstream/downstream services
   6. Known vulnerabilities or ongoing issues
   7. Runbook URLs and monitoring dashboards
   8. Past similar incidents and resolution patterns
   9. Team's Slack channel
   
   Output as JSON:
   {
     "service_name": "",
     "service_tier": "tier1|tier2|tier3",
     "sla_minutes": 0,
     "owners": [],
     "on_call_engineers": [],
     "slack_channel": "",
     "recent_deployments": [],
     "recent_changes": [],
     "dependencies": [],
     "runbook_url": "",
     "dashboard_url": "",
     "past_incidents_summary": "",
     "known_issues": [],
     "jira_project_key": ""
   }
   
   Rules:
   - Only include data that exists in Port
   - Use empty strings/arrays if not found

3. OpenAI 评估严重度与升级建议

   添加 OpenAI 节点（model: gpt-4o-mini），将 Port 上下文与事件标题、服务名、紧急程度一并送入。模型返回 JSON：严重度等级（critical/high/medium/low）、是否升级、升级原因、可能根因、推荐行动、调查清单。

   运行提示词 · 来源原文。OpenAI 节点 'Assess Severity' 的 chat messages[0].content 字段

   复制提示词

   =You are an incident management expert. Assess this incident and determine if immediate escalation is needed.
   
   Incident:
   - Title: {{ $('On PagerDuty Incident').item.json.body.event.data.title }}
   - Service: {{ $('On PagerDuty Incident').item.json.body.event.data.service.name }}
   - Urgency: {{ $('On PagerDuty Incident').item.json.body.event.data.urgency }}
   
   Context from Port:
   {{ $json.result.message }}
   
   Consider:
   1. Is this a production service?
   2. Were there recent deployments that could be related?
   3. Are there known issues affecting this service?
   4. What is the SLA requirement?
   
   Provide your response as JSON:
   {
     "severity_assessment": "critical|high|medium|low",
     "needs_escalation": true or false,
     "escalation_reason": "reason if needs_escalation is true",
     "likely_cause": "brief assessment of likely cause",
     "recommended_actions": ["action 1", "action 2", "action 3"],
     "investigation_checklist": ["check 1", "check 2", "check 3"]
   }
   
   Return ONLY the JSON object.

4. Slack 分级通知团队与领导层

   添加 If 节点判断 needs_escalation。无论是否升级，均向团队 Slack 频道发送事件详情、严重度、值班人、可能根因、推荐行动、近期变更及资源链接；若需升级，额外向 #leadership-alerts 发送升级通知。频道名从 Port 上下文读取，缺失则默认 #incidents。

   运行提示词 · 来源原文。Slack 节点 'Notify Team Channel' 的 text 字段

   复制提示词

   =:rotating_light: *INCIDENT ALERT*
   
   *{{ $('On PagerDuty Incident').item.json.body.event.data.title }}*
   
   *Service:* {{ $('Parse Context (New Incident)').item.json.result.message.parseJson().service_name }} ({{ $('Parse Context (New Incident)').item.json.result.message.parseJson().service_tier }})
   *Urgency:* {{ $('On PagerDuty Incident').item.json.body.event.data.urgency }}
   *Severity Assessment:* {{ $('Assess Severity').item.json.message.content.parseJson().severity_assessment }}
   
   *On-Call:* {{ $('Parse Context (New Incident)').item.json.result.message.parseJson().on_call_engineers.join(', ') || 'Check PagerDuty' }}
   
   *Likely Cause:*
   {{ $('Assess Severity').item.json.message.content.parseJson().likely_cause }}
   
   *Recommended Actions:*
   {{ $('Assess Severity').item.json.message.content.parseJson().recommended_actions.map((a, i) => (i+1) + '. ' + a).join('\n') }}
   
   *Recent Changes:*
   {{ $('Parse Context (New Incident)').item.json.result.message.parseJson().recent_deployments.slice(0,3).map(d => '- ' + d).join('\n') || 'None in last 48h' }}
   
   *Resources:*
   {{ $('Parse Context (New Incident)').item.json.result.message.parseJson().runbook_url ? '<' + $('Parse Context (New Incident)').item.json.result.message.parseJson().runbook_url + '|Runbook>' : '' }} {{ $('Parse Context (New Incident)').item.json.result.message.parseJson().dashboard_url ? '| <' + $('Parse Context (New Incident)').item.json.result.message.parseJson().dashboard_url + '|Dashboard>' : '' }}
   
   :pagerduty: <{{ $('On PagerDuty Incident').item.json.body.event.data.html_url }}|View in PagerDuty>

   运行提示词 · 来源原文。Slack 节点 'Escalate to Leadership' 的 text 字段

   复制提示词

   =:rotating_light::rotating_light: *INCIDENT ESCALATION*
   
   *{{ $('On PagerDuty Incident').item.json.body.event.data.title }}*
   
   *Service:* {{ $('Parse Context (New Incident)').item.json.result.message.parseJson().service_name }} ({{ $('Parse Context (New Incident)').item.json.result.message.parseJson().service_tier }})
   *Escalation Reason:* {{ $('Assess Severity').item.json.message.content.parseJson().escalation_reason }}
   
   *Team Channel:* {{ $('Parse Context (New Incident)').item.json.result.message.parseJson().slack_channel || '#incidents' }}
   *On-Call:* {{ $('Parse Context (New Incident)').item.json.result.message.parseJson().on_call_engineers.join(', ') || 'Check PagerDuty' }}
   
   :pagerduty: <{{ $('On PagerDuty Incident').item.json.body.event.data.html_url }}|View in PagerDuty>
   
   _Response team has been notified._

5. 创建 Jira 事件工单并通知

   添加 Jira 节点，项目键从 Port 上下文读取（默认 INC），工单类型 Bug，优先级按严重度映射（critical→Highest, high→High, 其他→Medium）。描述包含完整事件详情、服务上下文、AI 评估、调查清单、推荐行动、近期变更、资源链接及历史事件摘要。工单创建后向团队频道发送确认消息。

   运行提示词 · 来源原文。Slack 节点 'Notify Jira Created' 的 text 字段

   复制提示词

   =:jira: *Incident Ticket Created*
   
   {{ $json.key }}
   
   Track investigation and resolution in Jira.

6. 解决后提取上下文并生成事后分析模板

   事件解决后，Port 节点（agentIdentifier: context_retriever_agent）查询服务详情、负责人、Slack 频道、事件窗口内变更、Confluence/Notion 空间、日历集成信息，并计算 incident_duration_minutes。随后 OpenAI 节点（gpt-4o-mini）基于上下文生成结构化事后分析 JSON 模板，含时间线、影响、根因分析占位、行动项、经验教训。

   运行提示词 · 来源原文。Port 节点 'Extract Context (Resolved)' 的 prompt 字段

   复制提示词

   =A PagerDuty incident has been resolved.
   
   Incident Details:
   - ID: {{ $('On PagerDuty Incident').item.json.body.event.data.id }}
   - Title: {{ $('On PagerDuty Incident').item.json.body.event.data.title }}
   - Service: {{ $('On PagerDuty Incident').item.json.body.event.data.service.name }}
   - Created At: {{ $('On PagerDuty Incident').item.json.body.event.data.created_at }}
   - Resolved At: {{ $('On PagerDuty Incident').item.json.body.event.data.resolved_at || new Date().toISOString() }}
   
   Query the Port catalog for post-incident context.
   
   Retrieve:
   1. Service details and owners
   2. Team's Slack channel
   3. Changes made during the incident window
   4. Confluence/Notion space for post-mortems
   5. Calendar integration details for scheduling
   
   Calculate incident_duration_minutes from created_at to resolved_at.
   
   Output as JSON:
   {
     "service_name": "",
     "owners": [],
     "slack_channel": "",
     "confluence_space": "",
     "calendar_id": "",
     "incident_duration_minutes": 0,
     "changes_during_incident": [],
     "stakeholders_to_invite": []
   }
   
   Only include data that exists in Port.

   运行提示词 · 来源原文。OpenAI 节点 'Generate Post-Mortem Template' 的 chat messages[0].content 字段

   复制提示词

   =Create a post-mortem document template for this resolved incident.
   
   Incident:
   - ID: {{ $('On PagerDuty Incident').item.json.body.event.data.id }}
   - Title: {{ $('On PagerDuty Incident').item.json.body.event.data.title }}
   - Service: {{ $json.result.message.parseJson().service_name }}
   - Duration: {{ $json.result.message.parseJson().incident_duration_minutes }} minutes
   - Created: {{ $('On PagerDuty Incident').item.json.body.event.data.created_at }}
   - Resolved: {{ $('On PagerDuty Incident').item.json.body.event.data.resolved_at }}
   
   Context:
   {{ $json.result.message }}
   
   Create a structured post-mortem template as JSON:
   {
     "title": "Post-Mortem: [incident title]",
     "summary": "Brief 2-3 sentence summary",
     "timeline": [
       {"time": "HH:MM", "event": "Incident detected"},
       {"time": "HH:MM", "event": "Response initiated"},
       {"time": "HH:MM", "event": "Root cause identified"},
       {"time": "HH:MM", "event": "Resolution applied"},
       {"time": "HH:MM", "event": "Incident resolved"}
     ],
     "impact": "Description of customer/business impact",
     "root_cause_analysis": "[To be filled during post-mortem meeting]",
     "action_items": [
       {"action": "Suggested action 1", "owner": "TBD", "due_date": "TBD"},
       {"action": "Suggested action 2", "owner": "TBD", "due_date": "TBD"}
     ],
     "lessons_learned": "[To be filled during post-mortem meeting]",
     "meeting_attendees": []
   }
   
   Return ONLY the JSON object.

7. Port AI Agent 调度复盘会议并通知团队

   添加 Port 节点（agentIdentifier: post_mortem_scheduler_agent），输入事件详情、持续时间、已解决时间、利益相关者列表及事后分析模板内容。Agent 任务：创建事后分析文档、安排 2-3 个工作日内会议、邀请相关人员、在会议邀请中附加文档链接。返回 JSON 含文档创建状态、会议时间、受邀人列表。解析响应后向团队频道发送解决通知，包含 MTTR、文档链接和会议详情。

   运行提示词 · 来源原文。Port 节点 'Schedule Post-Mortem via Port' 的 prompt 字段

   复制提示词

   =An incident has been resolved and needs a post-mortem meeting scheduled.
   
   Incident Details:
   - ID: {{ $('On PagerDuty Incident').item.json.body.event.data.id }}
   - Title: {{ $('On PagerDuty Incident').item.json.body.event.data.title }}
   - Service: {{ $('Parse Context (Resolved)').item.json.result.message.parseJson().service_name }}
   - Duration: {{ $('Parse Context (Resolved)').item.json.result.message.parseJson().incident_duration_minutes }} minutes
   - Resolved At: {{ $('On PagerDuty Incident').item.json.body.event.data.resolved_at }}
   
   Stakeholders to invite:
   {{ $('Parse Context (Resolved)').item.json.result.message.parseJson().stakeholders_to_invite.join(', ') || $('Parse Context (Resolved)').item.json.result.message.parseJson().owners.join(', ') }}
   
   Post-Mortem Template:
   {{ $('Generate Post-Mortem Template').item.json.message.content }}
   
   Your tasks:
   1. Create a post-mortem document (in Confluence/Notion if available) with the template content
   2. Schedule a post-mortem meeting for 2-3 business days from now
   3. Invite relevant stakeholders (service owners, on-call engineers who responded)
   4. Include the incident details and post-mortem document link in the meeting invite
   
   Respond with JSON:
   {
     "document_created": true or false,
     "document_url": "URL if created",
     "meeting_scheduled": true or false,
     "meeting_time": "ISO timestamp",
     "attendees_invited": [],
     "status": "success|partial|failed",
     "message": "Summary of actions taken"
   }

   运行提示词 · 来源原文。Slack 节点 'Notify Incident Resolved' 的 text 字段

   复制提示词

   =:white_check_mark: *INCIDENT RESOLVED*
   
   *{{ $('On PagerDuty Incident').item.json.body.event.data.title }}*
   
   *Service:* {{ $('Parse Context (Resolved)').item.json.result.message.parseJson().service_name }}
   *Duration:* {{ $('Parse Context (Resolved)').item.json.result.message.parseJson().incident_duration_minutes }} minutes
   *MTTR:* {{ Math.floor($('Parse Context (Resolved)').item.json.result.message.parseJson().incident_duration_minutes / 60) }}h {{ $('Parse Context (Resolved)').item.json.result.message.parseJson().incident_duration_minutes % 60 }}m
   
   *Post-Mortem:*
   {{ $json.result.message.parseJson().document_url ? '<' + $json.result.message.parseJson().document_url + '|View Document>' : 'Document creation pending' }}
   
   *Post-Mortem Meeting:*
   {{ $json.result.message.parseJson().meeting_scheduled ? 'Scheduled for ' + $json.result.message.parseJson().meeting_time : 'Scheduling in progress' }}
   {{ $json.result.message.parseJson().attendees_invited.length > 0 ? 'Attendees: ' + $json.result.message.parseJson().attendees_invited.join(', ') : '' }}
   
   {{ $json.result.message.parseJson().message }}
   
   :pagerduty: <{{ $('On PagerDuty Incident').item.json.body.event.data.html_url }}|View in PagerDuty>

8. 回写 MTTR 指标到 Port

   添加 Port 节点（agentIdentifier: metrics_logger_agent），输入事件 ID、服务名、MTTR（分钟）、解决时间。任务：更新 Port 中服务实体的最新 MTTR、递增事件计数、更新最后事件时间戳。返回 JSON 确认指标是否已记录。

   运行提示词 · 来源原文。Port 节点 'Log MTTR to Port' 的 prompt 字段

   复制提示词

   =Log the incident resolution metrics to Port.
   
   Incident:
   - ID: {{ $('On PagerDuty Incident').item.json.body.event.data.id }}
   - Service: {{ $('Parse Context (Resolved)').item.json.result.message.parseJson().service_name }}
   - MTTR: {{ $('Parse Context (Resolved)').item.json.result.message.parseJson().incident_duration_minutes }} minutes
   - Resolved At: {{ $('On PagerDuty Incident').item.json.body.event.data.resolved_at }}
   
   Update the service entity in Port with:
   1. Latest incident MTTR
   2. Increment incident count
   3. Update last incident timestamp
   
   Respond with JSON:
   {
     "metrics_logged": true or false,
     "service_updated": "",
     "message": ""
   }

你正在搭建一个 n8n 工作流，用于自动化事件全生命周期管理。请先确认以下前提：
1. 你使用的是自托管 n8n 实例（模板标注不支持 Cloud）。
2. 已在 Port.io 注册并完成服务目录、值班表、部署历史的配置。
3. 已在 n8n 中安装 Port 自定义节点，并配置好 portIoApi、openAiApi、slackApi、jiraSoftwareCloudApi 四类凭证。
4. PagerDuty webhook 可访问你的 n8n 实例公网地址。
5. Slack 频道 #incidents 与 #leadership-alerts 已创建，Jira 项目键为 INC（或可自定义）。

确认后，按步骤导入 workflow.json 或按节点描述手动搭建，重点检查：
- Webhook 路径与 PagerDuty 配置一致；
- Switch 节点规则匹配 incident.triggered / incident.resolved；
- Port 节点的 agentIdentifier 与你在 Port 中配置的 AI Agent 名称一致；
- OpenAI 节点模型选择与你账户可用模型一致（模板使用 gpt-4o-mini）。

首次运行前，用 PagerDuty 的测试事件验证触发流，确认 Slack 能收到消息、Jira 能创建工单后，再测试解决流。