docs(memory): Uptime Kuma 接入 Bark 告警并通过端到端验证

.workbuddy/memory/2026-09-13.md

@@ -1,5 +1,6 @@
 # 2026-09-13
 
+2026-09-13 | CHG | Uptime Kuma 接入 Bark 告警：新增通知渠道「Bark 手机推送」（type 必须写 `Bark` 首字母大写，端点用容器网关 `http://172.17.0.1:8081/<key>`），已关联全部 9 个监控；用临时监控触发真实 DOWN 完成端到端验证（Bark 日志实测收到并返 200）后清理
 2026-09-13 | DEP | 安装用户级技能 `design-taste-frontend`（taste-skill v2，源 github.com/Leonxlnx/taste-skill，上游路径 skills/taste-skill/SKILL.md）；安装前完成安全审计（P2 安全，纯 Markdown 无脚本）；description 追加中文触发词；来源与审计记录见 `~/.workbuddy/skills/design-taste-frontend/SOURCE.md`
 2026-09-13 | FIX | 本机 git 走系统代理（HTTP_PROXY=127.0.0.1:58412）导致 github clone 失败（Empty reply / CONNECT 502）→ 清空代理 env 后用 `curl --noproxy "*"` 拉 codeload tarball 再解压可正常下载；另 Bash 工具在本机不可用（shim 报 dirname: command not found），需改用 PowerShell
 2026-09-13 | CHG | Uptime Kuma 第一批监控上线：9 个站点 HTTP 监控（60/120/300s 三档 + 证书到期通知），因官方 API 不支持创建监控，采用「直写 SQLite monitor 表 + 重启容器」批量添加

.workbuddy/memory/DEPLOY.md

@@ -67,5 +67,12 @@
   | file / read / tool / sub2api | 120s | 2 | 2000 |
   | code / status | 300s | 2 | 3000 |
 - **`service.xpcool.com` 单独放宽为 `["200-399","404"]`**：该后端是纯 POST 接口（遵循本项目"接口全 POST"规范），任何 GET 都 404；返回 404 说明 nginx + GoFrame 存活，**后端进程真挂时 nginx 返回 502，仍会判 DOWN**。
-- **通知渠道（Bark，2026-09-13 勘查结论）**：Kuma 2.5.3 **原生支持 Bark**（`/app/server/notification-providers/bark.js`），对应 `notification` 表（字段 `id/name/active/user_id/is_default/config`，`config` 为 JSON 字符串；截至 09-13 该表为空 = 尚未配通知）。Bark 服务端 = 容器 `bark`（`finab/bark-server`，`0.0.0.0:8081`），另经 `service.xpcool.com/bark/` 反代（末尾斜杠会剥掉 `/bark/` 前缀）；当前 server-monitor.sh 用的 key 在 `/etc/server-monitor.conf` 的 `BARK_URL`。**注意**：Kuma 容器访问 Bark 建议用宿主机地址（容器内 `127.0.0.1` 不是宿主机），即 `http://172.17.0.1:8081/<key>`（或把 kuma 加入 `xpcool-net` 后用容器名 `bark`）。
+- **通知渠道 = Bark（2026-09-13 已配置并验证）**：Kuma 2.5.3 原生支持 Bark（`/app/server/notification-providers/bark.js`）。`notification` 表：`id/name/active/user_id/is_default/config`，**`config` 是整条通知对象的 JSON 字符串（含 `type`），Kuma 用 `providerList[notification.type]` 派发，靠 `item.name` 索引**。
-- **待办**：「证书到期提前几天通知」需在面板设置里确认。
+  - **最大的坑：`type` 必须写 `Bark`（首字母大写）**，不是 `bark`。前端 `NotificationDialog.vue:257` 就是 `Bark: "Bark"`；写成小写则查不到 provider，**静默不发通知**。
+  - 当前配置：`id=1 / name='Bark 手机推送' / is_default=1`，`config` = `{"type":"Bark","barkEndpoint":"http://172.17.0.1:8081/sm5WnyANWx89GgLtFnRXhk","barkGroup":"Uptime-Kuma","apiVersion":"v1"}`。
+  - **端点必须用容器网关 `172.17.0.1:8081`**：kuma 在默认 `bridge` 网络（网关 172.17.0.1），容器内的 `127.0.0.1` 不是宿主机；Bark 容器端口发布在宿主机 `0.0.0.0:8081`，故走网关可达（实测 200）。
+  - `apiVersion:"v1"` → 走 GET `{endpoint}/{title}/{body}?icon=…&group=…&sound=…`（bark-server 支持）；不填也默认 v1。Bark key 就是 `/etc/server-monitor.conf` 里 `BARK_URL` 那个（`sm5WnyANWx89GgLtFnRXhk`），手机端就是 `https://service.xpcool.com/bark/<key>/`。
+  - 挂到监控：`monitor_notification(monitor_id, notification_id)`；**等价于 UI 的 `applyExisting`**。已为 9 个监控全部关联（共 9 行）。
+  - 脚本：`.workbuddy/tmp/kuma-bark-setup.sh`（幂等写入）、`.workbuddy/tmp/kuma-bark-verify.sh`（端到端验证 + 自清理）；改库前必备份 `/tmp/kuma.db.backup.<时间戳>`。
+  - **端到端验证方法（推荐复用）**：临时插一个指向 `http://127.0.0.1:9/` 的监控（`interval=20/retry_interval=20/maxretries=0`）→ 重启容器 → 十几秒即 DOWN 并推送 → 查 `docker logs bark` 确认真的收到 GET 推送（200）→ 删掉临时监控与其 heartbeat 再重启。全程不清真站点，可靠性最高。
+  - 全部 9 个监控 `expiry_notification=1`（证书到期通知已开），当前均 UP。

.workbuddy/tmp/kuma-bark-inspect.sh

@@ -0,0 +1,37 @@
+#!/bin/bash
+# 1) 确认 Bark key 可用（发一条真实测试推送）  2) 取 Kuma Bark 通知配置字段  3) 查容器网络（决定用哪个地址访问 Bark）
+# 2026-09-13
+
+KEY="sm5WnyANWx89GgLtFnRXhk"
+
+echo "=== 1. Bark 服务自检 ==="
+curl -s -o /dev/null -m 10 -w "healthz=%{http_code}\n" http://127.0.0.1:8081/healthz
+
+echo ""
+echo "=== 2. 发送真实测试推送（宿主机 → Bark） ==="
+curl -s -m 20 -o /tmp/bark-test.json -w "http=%{http_code}\n" \
+  "http://127.0.0.1:8081/$KEY/Kuma%20Test/ai-verify-20260913"
+echo "--- resp ---"
+cat /tmp/bark-test.json 2>/dev/null
+echo ""
+
+echo ""
+echo "=== 3. Kuma 容器网络 ==="
+sudo docker inspect uptime-kuma --format '{{json .NetworkSettings.Networks}}'
+echo ""
+echo "--- 容器内能否访问宿主机 Bark (172.17.0.1) ---"
+sudo docker exec uptime-kuma sh -c "curl -s -o /dev/null -m 8 -w 'code=%{http_code}\n' http://172.17.0.1:8081/healthz" 2>&1 || echo "container curl 不可用"
+
+echo ""
+echo "=== 4. Bark 通知 provider 源码（取字段名） ==="
+sudo docker exec uptime-kuma cat /app/server/notification-providers/bark.js 2>&1 | head -60
+
+echo ""
+echo "=== 5. Kuma 表结构 ==="
+sudo docker exec uptime-kuma sqlite3 /app/data/kuma.db "PRAGMA table_info(notification);" 2>&1
+echo "--- monitor_notification ---"
+sudo docker exec uptime-kuma sqlite3 /app/data/kuma.db "PRAGMA table_info(monitor_notification);" 2>&1
+echo "--- 现有通知 ---"
+sudo docker exec uptime-kuma sqlite3 /app/data/kuma.db "SELECT id,name,active,user_id,is_default FROM notification;" 2>&1
+echo "--- 现有监控 ---"
+sudo docker exec uptime-kuma sqlite3 /app/data/kuma.db "SELECT id,name FROM monitor ORDER BY id;" 2>&1

.workbuddy/tmp/kuma-bark-setup.sh

@@ -0,0 +1,58 @@
+#!/bin/bash
+# 为 Uptime Kuma 配置 Bark 通知渠道，并挂到全部监控（2026-09-13）
+# 幂等：重复执行不会产生重复行
+set -e
+
+DB=/app/data/kuma.db
+KEY="sm5WnyANWx89GgLtFnRXhk"
+KCONF="/data/uptime-kuma/kuma.db"
+TS=$(date +%Y%m%d%H%M%S)
+
+echo "=== 0. 备份数据库 ==="
+sudo cp "$KCONF" "/tmp/kuma.db.backup.$TS"
+ls -l "/tmp/kuma.db.backup.$TS"
+
+echo ""
+echo "=== 1. 用 Kuma 的 URL 格式在容器内实测一次（含 icon/group/sound 参数） ==="
+sudo docker exec uptime-kuma sh -c "curl -s -o /dev/null -m 15 -w 'kuma_style_url=%{http_code}\n' 'http://172.17.0.1:8081/$KEY/UptimeKuma%20Monitor%20Down/test-from-kuma-container?icon=https%3A%2F%2Fgithub.com%2Flouislam%2Fuptime-kuma%2Fraw%2Fmaster%2Fpublic%2Ficon.png&group=Uptime-Kuma&sound=telegraph'"
+
+echo ""
+echo "=== 2. 写入 Bark 通知（若不存在） ==="
+NAME='Bark 手机推送'
+CONFIG='{"type":"Bark","name":"Bark 手机推送","barkEndpoint":"http://172.17.0.1:8081/sm5WnyANWx89GgLtFnRXhk","barkGroup":"Uptime-Kuma","apiVersion":"v1","isDefault":true,"active":true,"applyExisting":false}'
+
+EXIST=$(sudo docker exec uptime-kuma sqlite3 "$DB" "SELECT COUNT(*) FROM notification WHERE name='$NAME';")
+if [ "$EXIST" = "0" ]; then
+  sudo docker exec uptime-kuma sqlite3 "$DB" "INSERT INTO notification (name, active, user_id, is_default, config) VALUES ('$NAME', 1, 1, 1, '$CONFIG');"
+  echo "已新增通知渠道"
+else
+  sudo docker exec uptime-kuma sqlite3 "$DB" "UPDATE notification SET active=1, is_default=1, config='$CONFIG' WHERE name='$NAME';"
+  echo "已更新既有通知渠道"
+fi
+
+echo "--- JSON 合法性校验 ---"
+sudo docker exec uptime-kuma sqlite3 "$DB" "SELECT id, json_valid(config) AS json_ok, json_extract(config,'\$.type') AS type, json_extract(config,'\$.barkEndpoint') AS endpoint FROM notification;"
+
+NID=$(sudo docker exec uptime-kuma sqlite3 "$DB" "SELECT id FROM notification WHERE name='$NAME' LIMIT 1;")
+echo "notification_id=$NID"
+
+echo ""
+echo "=== 3. 挂到全部监控（幂等） ==="
+sudo docker exec uptime-kuma sqlite3 "$DB" "INSERT INTO monitor_notification (monitor_id, notification_id) SELECT m.id, $NID FROM monitor m WHERE NOT EXISTS (SELECT 1 FROM monitor_notification mn WHERE mn.monitor_id = m.id AND mn.notification_id = $NID);"
+echo "--- 关联结果 ---"
+sudo docker exec uptime-kuma sqlite3 "$DB" "SELECT mn.monitor_id, m.name, mn.notification_id FROM monitor_notification mn JOIN monitor m ON m.id = mn.monitor_id ORDER BY mn.monitor_id;"
+
+echo ""
+echo "=== 4. 重启容器加载配置 ==="
+sudo docker restart uptime-kuma > /dev/null
+sleep 15
+sudo docker ps --filter name=uptime-kuma --format '{{.Names}} | {{.Status}}'
+curl -s -o /dev/null -m 10 -w "panel_http=%{http_code}\n" http://127.0.0.1:3001/
+
+echo ""
+echo "=== 5. 复核 ==="
+sudo docker exec uptime-kuma sqlite3 "$DB" "SELECT id,name,active,is_default FROM notification;"
+sudo docker exec uptime-kuma sqlite3 "$DB" "SELECT COUNT(*) AS linked FROM monitor_notification;"
+echo "=== 容器日志尾部（看有无通知相关报错） ==="
+sudo docker logs --tail 20 uptime-kuma 2>&1 | grep -i -E "error|notification|bark" || echo "(无相关日志)"
+echo "DONE"

.workbuddy/tmp/kuma-bark-src.sh

@@ -0,0 +1,12 @@
+#!/bin/bash
+# 查看 Kuma Bark 通知的完整实现与 config 结构要求
+echo "=== bark.js 全文 ==="
+sudo docker exec uptime-kuma cat /app/server/notification-providers/bark.js 2>&1
+
+echo ""
+echo "=== notification-provider.js 的 postNotification ==="
+sudo docker exec uptime-kuma sed -n '1,80p' /app/server/notification-providers/notification-provider.js 2>&1
+
+echo ""
+echo "=== notification.js 里 config / type 的解析 ==="
+sudo docker exec uptime-kuma grep -n "JSON.parse\|notification.type\|config" /app/server/notification.js 2>&1 | head -30

.workbuddy/tmp/kuma-bark-verify.sh

@@ -0,0 +1,69 @@
+#!/bin/bash
+# 端到端验证 Bark 告警：临时加一个必然失败的监控 → 触发 DOWN → 看推送是否送达 → 清理
+# 2026-09-13
+
+DB=/app/data/kuma.db
+Q() { sudo docker exec uptime-kuma sqlite3 "$DB" "$1"; }
+KCONF=/data/uptime-kuma/kuma.db
+TS=$(date +%Y%m%d%H%M%S)
+
+sudo cp "$KCONF" "/tmp/kuma.db.backup.$TS"
+
+echo "=== 1. 新增临时监控（指向必然连不上的端口） ==="
+Q "INSERT INTO monitor (name, active, user_id, type, url, interval, retry_interval, maxretries, weight, accepted_statuscodes_json, description) VALUES ('自检 · 临时（验证后自动删除）', 1, 1, 'http', 'http://127.0.0.1:9/', 20, 20, 0, 9999, '[\"200-399\"]', 'Bark 通道端到端验证');"
+
+TMPID=$(Q "SELECT id FROM monitor WHERE name='自检 · 临时（验证后自动删除）' LIMIT 1;")
+NID=$(Q "SELECT id FROM notification WHERE name='Bark 手机推送' LIMIT 1;")
+echo "temp_monitor_id=$TMPID  notification_id=$NID"
+
+Q "INSERT INTO monitor_notification (monitor_id, notification_id) VALUES ($TMPID, $NID);"
+
+echo ""
+echo "=== 2. 重启 Kuma 让它加载临时监控并开始探测 ==="
+sudo docker restart uptime-kuma > /dev/null
+sleep 20
+sudo docker ps --filter name=uptime-kuma --format '{{.Names}} | {{.Status}}'
+
+echo ""
+echo "=== 3. 等待探测与告警推送（约 60 秒） ==="
+for i in 1 2 3 4 5 6; do
+  sleep 15
+  ST=$(Q "SELECT status FROM heartbeat WHERE monitor_id=$TMPID ORDER BY id DESC LIMIT 1;" 2>/dev/null)
+  echo "  [${i}] 第 $((i*15))s：heartbeat status=${ST:-无}"
+  if [ "$ST" = "0" ]; then break; fi
+done
+
+echo ""
+echo "=== 4. 该临时监控的心跳记录（0=DOWN） ==="
+Q "SELECT id, status, msg, time FROM heartbeat WHERE monitor_id=$TMPID ORDER BY id DESC LIMIT 5;"
+
+echo ""
+echo "=== 5. 通知发送历史 ==="
+Q "SELECT name FROM sqlite_master WHERE type='table' AND name LIKE '%notification%';"
+Q "SELECT * FROM notification_sent_history ORDER BY id DESC LIMIT 5;" 2>/dev/null || echo "(无该表或为空)"
+
+echo ""
+echo "=== 6. Bark 服务端日志（能看到实际收到的推送） ==="
+sudo docker logs --tail 25 bark 2>&1 | tail -25
+
+echo ""
+echo "=== 7. Kuma 日志中的通知相关行 ==="
+sudo docker logs --tail 60 uptime-kuma 2>&1 | grep -i -E "bark|notification|error" | tail -15 || echo "(无)"
+
+echo ""
+echo "=== 8. 清理临时监控 ==="
+Q "DELETE FROM heartbeat WHERE monitor_id=$TMPID;"
+Q "DELETE FROM monitor_notification WHERE monitor_id=$TMPID;"
+Q "DELETE FROM monitor WHERE id=$TMPID;"
+echo "--- 清理后剩余监控 ---"
+Q "SELECT id, name FROM monitor ORDER BY id;"
+echo "--- 清理后关联数（应为 9） ---"
+Q "SELECT COUNT(*) FROM monitor_notification;"
+
+echo ""
+echo "=== 9. 重启并复核 ==="
+sudo docker restart uptime-kuma > /dev/null
+sleep 15
+sudo docker ps --filter name=uptime-kuma --format '{{.Names}} | {{.Status}}'
+curl -s -o /dev/null -m 10 -w "panel_http=%{http_code}\n" http://127.0.0.1:3001/
+echo "DONE"

.workbuddy/tmp/kuma-final-check.sh

@@ -0,0 +1,19 @@
+#!/bin/bash
+# 复核：Bark 通知 + 各监控的证书到期通知开关
+DB=/app/data/kuma.db
+Q() { sudo docker exec uptime-kuma sqlite3 "$DB" "$1"; }
+
+echo "=== 通知渠道 ==="
+Q "SELECT id, name, active, is_default, json_extract(config,'\$.type') AS type, json_extract(config,'\$.barkEndpoint') AS endpoint, json_extract(config,'\$.barkGroup') AS grp FROM notification;"
+
+echo ""
+echo "=== 监控 × 证书到期通知开关 ==="
+Q "SELECT id, name, expiry_notification, interval, maxretries FROM monitor ORDER BY id;"
+
+echo ""
+echo "=== 关联数 ==="
+Q "SELECT COUNT(*) FROM monitor_notification;"
+
+echo ""
+echo "=== 当前各监控状态（最近一条心跳） ==="
+Q "SELECT m.id, m.name, h.status, h.time FROM monitor m LEFT JOIN heartbeat h ON h.id = (SELECT id FROM heartbeat WHERE monitor_id = m.id ORDER BY id DESC LIMIT 1) ORDER BY m.id;"

.workbuddy/tmp/kuma-type.sh

@@ -0,0 +1,7 @@
+#!/bin/bash
+# 确认 providerList 的 key 命名（notification.type 该填什么）
+echo "=== providerList 构造 ==="
+sudo docker exec uptime-kuma grep -n -B3 -A12 "providerList" /app/server/notification.js 2>&1 | head -50
+echo ""
+echo "=== providers 列表引入 ==="
+sudo docker exec uptime-kuma grep -n -A40 "require(\"./notification-providers" /app/server/notification.js 2>&1 | head -50

.workbuddy/tmp/kuma-type2.sh

@@ -0,0 +1,16 @@
+#!/bin/bash
+# 确认前端提交的 notification.type 到底是 "bark" 还是 "Bark"
+echo "=== NotificationDialog 中 type 的来源 ==="
+sudo docker exec uptime-kuma ls /app/src/components/notifications/ 2>&1 | head -30
+echo ""
+echo "=== 搜 bark 类型常量 ==="
+sudo docker exec uptime-kuma grep -rn -i "bark" /app/src/components/NotificationDialog.vue /app/src/components/notifications/Bark.vue 2>&1 | head -20
+echo ""
+echo "=== 搜 type 列表定义（notification type 下拉） ==="
+sudo docker exec uptime-kuma grep -rn -i "\"bark\"\|'bark'\|value=\"bark\"" /app/src 2>&1 | head -20
+echo ""
+echo "=== 看 init() 里 providerList 赋值的完整上下文（是否 toLowerCase） ==="
+sudo docker exec uptime-kuma sed -n '220,245p' /app/server/notification.js 2>&1
+echo ""
+echo "=== 新增/保存通知时是否校验 type ==="
+sudo docker exec uptime-kuma sed -n '275,300p' /app/server/notification.js 2>&1