@@ -54,7 +54,9 @@ One integration can follow several status pages. Alerts from different pages are
 
 <Step title="Verify">
 
-Instatus has no button that sends a test notification. To verify:
+Saving the subscriber sends a validation delivery right away (the **RUN** button next to the Webhook URL field sends another). Flashduty recognizes it and answers with success, no alert.
+
+To verify with a real incident instead:
 
 1. In Instatus, go to **Incidents** and click **Add incident**. Set the status to **Investigating**, select an affected component, set it to **Major outage**, then save and notify subscribers. Flashduty shows an incident alert and a component alert, both Critical
 2. Add an update to that incident with the status **Resolved**, set the component back to **Operational**, and notify subscribers. Both alerts recover
@@ -79,7 +81,8 @@ Instatus sends one JSON payload each time an incident is added or updated, a com
 | `incident.id` | Incident ID | Alert Key, label `incident_id` |
 | `incident.name` | Incident name | Alert title |
 | `incident.status` | Incident status | Alert status, label `incident_status` |
-| `incident.impact` | Incident impact | Alert severity, label `impact` |
+| `incident.impact` | Incident impact | Label `impact`, informational only: on an incident created through the Instatus dashboard this carries the same word as `status` (e.g. `Investigating`), not an outage severity |
+| `incident.affected_components[].status` | Status of each affected component | Alert severity — the worst one across all affected components; used as a fallback only when there are none |
 | `incident.url` | Incident link | Label `incident_url` |
 | `incident.incident_updates` | Incident updates | The body of the latest update is the alert description, truncated above 8 KB |
 
@@ -122,15 +125,17 @@ Renaming an incident or component, or changing the impact, does not change the A
 
 **Instatus incidents**
 
-The alert severity comes from the impact (`impact`, which takes the same values as component status), and the alert status from the incident status (`status`):
+The alert severity is the worst status across the incident's `affected_components` (same values as component status, only differently cased and spaced, e.g. `Major outage`), and the alert status comes from the incident status (`status`):
 
-| Instatus `impact` | Flashduty severity |
+| Instatus `affected_components[].status` | Flashduty severity |
 | :--- | :--- |
-| `MAJOROUTAGE` | Critical |
-| `PARTIALOUTAGE` | Warning |
-| `DEGRADEDPERFORMANCE`, `OPERATIONAL` | Info |
+| `Major outage` | Critical |
+| `Partial outage` | Warning |
+| `Degraded performance`, `Operational` | Info |
 | Empty or any other value | Warning |
 
+An incident with no affected component falls back to the same mapping applied to `impact` instead — but on a dashboard-created incident, `impact` is usually just the `status` wording (e.g. `Investigating`), which matches none of these values and lands on Warning.
+
 | Instatus `status` | Flashduty status |
 | :--- | :--- |
 | `INVESTIGATING`, `IDENTIFIED`, `MONITORING` | Trigger or update the alert |
 
@@ -54,9 +54,11 @@ keywords: ["告警集成", "Instatus", "状态页", "第三方服务状态", "We
 
 <Step title="验证">
 
-Instatus 没有发送测试通知的按钮。可以这样验证：
+保存订阅者时，Instatus 会立即发送一次校验推送（Webhook URL 输入框旁的 **RUN** 按钮可以重新触发）。Flashduty 会识别这次校验推送并返回成功、不生成告警。
 
-1. 在 Instatus 中进入 **Incidents**，点击 **Add incident**，状态选 **Investigating**，勾选一个受影响的组件并把它的状态设为 **Major outage**，保存并通知订阅者。Flashduty 中出现一条事件告警和一条组件告警，等级为 Critical
+也可以驱动一次真实的事件来验证：
+
+1. 在 Instatus 中进入 **Incidents**，点击 **Add incident**，状态选 **Investigating**，勾选一个受影响的组件并把它的状态设为 **Major outage**，保存并通知订阅者。Flashduty 中出现一条事件告警和一条组件告警，等级都是 Critical
 2. 为这个事件添加一条更新，状态选 **Resolved**，组件恢复为 **Operational**，并通知订阅者。两条告警随之恢复
 
 发布事件或更新时如果没有通知订阅者，Instatus 不会推送，Flashduty 也不会收到。
@@ -79,7 +81,8 @@ Instatus 在事件新增或更新、组件状态变化、计划维护新增或
 | `incident.id` | 事件 ID | Alert Key，标签 `incident_id` |
 | `incident.name` | 事件名称 | 告警标题 |
 | `incident.status` | 事件状态 | 告警状态，标签 `incident_status` |
-| `incident.impact` | 事件影响程度 | 告警等级，标签 `impact` |
+| `incident.impact` | 事件影响程度 | 标签 `impact`，仅作参考：通过 Instatus 控制台创建的事件，这个字段的取值和 `status` 相同（如 `Investigating`），并不是故障等级 |
+| `incident.affected_components[].status` | 受影响组件的状态 | 告警等级，取受影响组件中最高的等级；没有受影响组件时才退回使用 `impact` |
 | `incident.url` | 事件详情链接 | 标签 `incident_url` |
 | `incident.incident_updates` | 事件的更新记录 | 最新一条更新的正文作为告警描述，超过 8 KB 时截断 |
 
@@ -122,15 +125,17 @@ Flashduty 按推送类型分别生成 Alert Key：
 
 **事件**
 
-事件的告警等级取自影响程度（`impact`，取值与组件状态相同），告警状态取自事件状态（`status`）：
+事件的告警等级取自受影响组件中最高的等级（`affected_components[].status`，取值和组件状态相同，只是大小写和间距不同，如 `Major outage`），告警状态取自事件状态（`status`）：
 
-| Instatus `impact` | Flashduty 告警等级 |
+| Instatus `affected_components[].status` | Flashduty 告警等级 |
 | :--- | :--- |
-| `MAJOROUTAGE` | Critical |
-| `PARTIALOUTAGE` | Warning |
-| `DEGRADEDPERFORMANCE`、`OPERATIONAL` | Info |
+| `Major outage` | Critical |
+| `Partial outage` | Warning |
+| `Degraded performance`、`Operational` | Info |
 | 空值或其他值 | Warning |
 
+事件没有受影响组件时，退回使用 `impact` 按同样的映射表取等级；但通过控制台创建的事件，`impact` 通常只是 `status` 的文字（例如 `Investigating`），映射不到任何已知取值，等级会落到 Warning。
+
 | Instatus `status` | Flashduty 状态 |
 | :--- | :--- |
 | `INVESTIGATING`、`IDENTIFIED`、`MONITORING` | 触发或更新告警 |