wooo/awoooi
1
1
Fork 0
Code Issues 2 Pull Requests 1 Actions 5 Packages Projects Releases Wiki Activity
1,082 Commits 195 Branches 2 Tags
e2c6ca598e4f641b5bb8867b1080ffbfd78e65dc
Commit Graph

305 Commits

Author SHA1 Message Date
OG T
e2c6ca598e fix(approval_db): update_telegram_message 用 raw SQL + CAST BIGINT 避免 int32 overflow
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
telegram_chat_id 為 BIGINT (5619078117 > 2^31-1)，SQLAlchemy ORM 會推斷為 $N::INTEGER
改用 raw SQL + CAST(:telegram_chat_id AS BIGINT) 繞過型別推斷

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 00:53:50 +08:00
OG T
a303b5ef91 feat(chat): NemoClaw 改接 Ollama 111 deepseek-r1:14b
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 4m6s
Details 
2026-04-09 ogt: 棄用 Claude Haiku，改用本地 deepseek-r1:14b
- 端點: http://192.168.0.111:11434
- 過濾 <think>...</think> 推理區塊，只回傳結論
- timeout 120s（14b 推理較慢）
- 完全免費，不計入 Claude API 費用

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 23:38:57 +08:00
OG T
62cb274735 feat(host_aggregator+k8s): 新增 121 K3s Worker 主機監控
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
HOST_CONFIGS 加入 192.168.0.121(K3s Worker):
- K3s API tcp:6443
- awoooi-api NodePort tcp:32334
- awoooi-web NodePort tcp:32335

NetworkPolicy 補開 121 egress: 6443/32334/32335
NodePort 服務實際在 121(mon1)，非 120(mon)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 23:36:36 +08:00
OG T
fc9d0f9c1f fix(host_aggregator): total=1 時 total//2=0 導致服務全 up 仍顯示 unhealthy
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
112(Kali) 和 120(K3s) 各只有 1 個服務，down_count=0 >= total//2=0
永遠成立 → 永遠 unhealthy。改為 total > 1 才套用 >=half 門檻。

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 23:35:37 +08:00
OG T
d324dd7aed fix(telegram): 移除所有告警訊息欄位截斷限制，放寬至 Telegram 4096 字元硬限
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
問題：root_cause[:50/100]、suggested_action[:80]、suggestion[:50]、
note[:150]、fix_action[:100]、impact[:150]、hypothesis[:200]
以及 message[:900]/[:1000] 導致告警內容顯示不完整。

修復：移除欄位截斷，整體上限改為 4096（Telegram API 硬限制）。

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 23:32:51 +08:00
OG T
eb46079b4a fix(telegram): root_cause 顯示長度 50→100 字元，符合 SOUL.md 鐵律
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
SOUL.md 明定根因摘要上限 100 字元，但程式碼兩處 IncidentApprovalCard
均截在 [:50]，導致告警卡片訊息被截斷。

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 23:30:58 +08:00
OG T
b1e207ffae fix(host_aggregator): E2E驗證後修正 HOST_CONFIGS — Ollama位置+NodePort+Nginx
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
從 K3s Pod 內 Python socket 實測確認後修正:
- 110: 加 Prometheus(9090) Grafana(3002)，移除 GH Runner(3000 refused)
- 112: 移除 SSH:22 (K3s Pod NetworkPolicy 未開)
- 120: 移除 awoooi NodePort(只在121不在120)
- 188: 移除 Ollama(在111非188) 和 Nginx:443(Pod內打不通)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 20:27:46 +08:00
OG T
21567a7a6d fix(host_aggregator): 修正四台主機 probe 端點錯誤導致全部顯示 unhealthy
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 12m1s
Details 
- 110: Harbor http→tcp(5000), Docker 2375→Gitea tcp(3001)
- 120: K3s 6443 https(401誤判)→tcp, 移除 Traefik 80(closed)
- 188: OpenClaw 8089→8088 (實際端口)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 19:52:34 +08:00
OG T
34f0228d92 fix(executor): K8s ClusterIP 10.43.0.1 不可達 — 加 K8S_API_SERVER_URL 覆蓋 + migration job
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 12m0s
Details 
問題: in-cluster config 讀到 10.43.0.1:443，但 K3s Pod 內 iptables/kube-proxy
      沒把流量導到實際 API server，導致 Connection refused，批准後 kubectl 永遠失敗

修復:
- executor.py: load_incluster_config() 後讀 K8S_API_SERVER_URL env 覆蓋 host
- 04-configmap.yaml: 設 K8S_API_SERVER_URL=https://192.168.0.120:6443
- migrate-sprint5r-telegram-message-id.yaml: approval_records 新增兩欄 migration job

E2E 驗證: kubectl rollout restart deployment/awoooi-worker success=True ✅

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 19:10:27 +08:00
OG T
ebccb88278 fix(approval_db): 修復 incident_id 篩選查空 DB 欄位而非 JSON 導致執行斷路
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
get_all_approvals(incident_id=...) 原本在應用層過濾
a.metadata.get("incident_id")，但 ApprovalRecord.incident_id
是直接欄位，不在 extra_metadata JSON，導致永遠返回空列表，
Telegram 批准後出現 telegram_approval_not_found_by_incident，
審批從未實際執行。改為 .where(ApprovalRecord.incident_id == incident_id)
DB 層直接篩選，同時效能更佳。

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 19:05:48 +08:00
OG T
896bef94ee fix(web): pending-approvals-card 加防重複點擊 + loading 狀態 
linter 自動強化: actioningId state 防止同一張卡重複操作
- disabled + opacity 0.6 + cursor not-allowed
- loading 時按鈕顯示 '...'
- finally() 確保 actioningId 清除

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 18:38:08 +08:00
OG T
890e2a9568 fix(review): 架構審查修復 — P0 import crash + i18n 零 hardcode + 靜默錯誤 
P0:
- proposal_service.py: 補 get_redis + INCIDENT_KEY_PREFIX import
  (修前: resolve_incident_after_approval 必 NameError crash)

P1 i18n:
- page.tsx: 拓撲群組移除 emoji，改用 tTopo() i18n key
- page.tsx: 主機標籤 (DevOps金庫等) 改 tTopo() i18n
- ai-model-status.tsx: 加 useTranslations，AI 模型狀態 → t('aiModelStatus')
- disposition-mini.tsx: 查看完整報表 → t('viewAllReport')
- recent-activity.tsx: 查看活動串流 → t('viewAllAlerts')

P2 品質:
- pending-approvals-card.tsx: approve/reject 加 r.ok 檢查+錯誤顯示，查看全部授權加路由+i18n
- page-tabs.tsx: TabSkeleton 載入中... → t('loading')
- page.tsx: ↑5% → tDashboard('trendUp', {pct}) 動態值
- page.tsx: Prometheus '23' hardcode → '-- targets'

i18n 新增 key (zh-TW + en 同步):
- dashboard: viewAllAlerts/viewAllAuth/viewAllReport/aiModelStatus/loading/trendUp
- topology: groupExternal/allReachable/investigating/hostDevops/hostAiData/hostK3sMaster/hostK3sWorker

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 18:34:50 +08:00
OG T
1483218bab feat(approval): 批准/拒絕後立即回應 Telegram + 持久化 message_id 到 DB
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 13m9s
Details 
問題：按下 TG 批准/拒絕按鈕後完全沒有任何回應，使用者不知道是否成功。
      Telegram message_id 只存 Redis 24h TTL，過期後無法追蹤。

修正：
- approval_records 加 telegram_message_id / telegram_chat_id 欄位（已 ALTER TABLE）
- approval_db.update_telegram_message() — 持久化 message_id 到 DB
- decision_manager: 發送告警卡片後同時寫 Redis + DB
- telegram_gateway._notify_approval_result() — 批准/拒絕後：
    1. editMessageReplyMarkup 移除批准/拒絕按鈕，保留資訊按鈕
    2. sendMessage reply_to 在原訊息下回覆狀態行
    3. fallback: send_notification 發新訊息
- _handle_group_command: chat_id 改為 _chat_id 消除 IDE lint

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 18:19:31 +08:00
OG T
2c7d5d049c fix(openclaw): Nemotron tool call 回填 kubectl_command，讓批准後執行器能解析
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
根本問題：Nemotron 產生的 restart_deployment(deployment_name=sentry) tool call
只存在 nemotron_tools[]，沒有回填到 proposal["kubectl_command"]。
proposal_service 拿到的 kubectl_command 是空的，approval_records.action 存空值，
parse_operation_from_action 永遠返回 None，execute_approved_action 永遠 skip。

修正：Nemotron (和 Gemini fallback) 成功後，將 tool call 轉換為 kubectl 指令
並回填 proposal["kubectl_command"]，讓 proposal_service 能取到可執行指令。

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 18:15:01 +08:00
OG T
ae9780837d fix(proposal): action 優先用 kubectl_command，修復批准後永遠 skip 執行的根本 bug 
根本問題：approval_records.action 存的是 LLM action_title（中文標題，如「重啟 sentry 服務」），
parse_operation_from_action() 無法解析，導致 execute_approved_action() 每次都 skip。

修正：action 優先取 llm_proposal["kubectl_command"]（可執行的 kubectl 指令），
僅在沒有 kubectl_command 時才 fallback 到 action_title。

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 18:13:22 +08:00
OG T
d8c2969341 feat(telegram): AI 鏈路透明化 — 告警訊息顯示 OpenClaw + Tool Calling 模型/後端
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 12m12s
Details 
- nemotron.py: 偵測 OllamaToolProvider vs NvidiaProvider，記錄 tool_model/tool_backend
- openclaw.py: 傳播 nemotron_tool_model/nemotron_tool_backend 到 proposal
- decision_manager.py: 從 proposal_data 提取並傳給 send_approval_card()
- telegram_gateway.py: TelegramMessage 新增兩個欄位，format_with_nemotron 顯示
  "🔧 Tool Calling: llama3.1:8b (Ollama 本機)" 或 "NVIDIA 雲端"

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 15:05:16 +08:00
OG T
7857c25677 feat: Ollama 本機 Tool Calling 取代 NVIDIA 雲端 (44s→~5s)
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
- nvidia_provider.py: 新增 OllamaToolProvider
  - 實作 INvidiaProvider protocol，打 Ollama /v1/chat/completions
  - 模型: llama3.1:8b (tool calling 最穩定的 8B)
  - 延遲: 44s → ~5s（本機 M1 Pro 192.168.0.111）
  - get_nvidia_provider() 根據 USE_OLLAMA_TOOL_CALLING 切換
- config.py: USE_OLLAMA_TOOL_CALLING=True (預設開啟), OLLAMA_TOOL_MODEL=llama3.1:8b
- 回退: USE_OLLAMA_TOOL_CALLING=False → 恢復 NvidiaProvider 雲端

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 14:55:04 +08:00
OG T
4f80ba38c0 feat: 告警狀態變更在原訊息延續 (方案 B)
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 12m28s
Details 
**telegram_gateway.py**
- 新增 append_incident_update(incident_id, status_line)
  - 從 Redis tg_msg:{id} 取 message_id
  - editMessageReplyMarkup: 移除 Row1(批准/拒絕/靜默)，保留 Row2(詳情/重診/歷史)
  - sendMessage reply_to_message_id: 在原訊息下方追加狀態行
  - 找不到 message_id 回傳 False（呼叫方自行 fallback）

**decision_manager.py**
- _push_decision_to_telegram: send_approval_card 後存 tg_msg:{id}=message_id (TTL 24h)
- _push_auto_repair_result: 改用 append_incident_update，找不到 message_id 才 fallback 新訊息

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 14:21:33 +08:00
OG T
2554ac1e60 fix: E2E test 告警識別 + 自動修復結果 Telegram 通知
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
**alert_rule_engine.py**
- _matches() 加入 instance_prefix 匹配（最高優先）
- match_rule() 傳入 instance label 至 _matches
- 用途: e2e-final-* / e2e-test-* instance 可被 YAML 規則識別

**alert_rules.yaml**
- 新增 e2e_smoke_test 規則 (priority=120)
- alertname: E2E_SMOKE_TEST / instance_prefix: e2e-final- / e2e-test- / test-host
- suggested_action: NO_ACTION，顯示「告警鏈路驗證成功」

**decision_manager.py**
- _auto_execute() 成功後發 Telegram 結果通知 ✅
- _auto_execute() 失敗後發 Telegram 失敗通知 ❌
- 新增 _push_auto_repair_result() 函數

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 14:16:15 +08:00
OG T
fc03eb1f4d fix(auto-repair): _extract_symptoms 優先用 labels.alertname 取得原始 alertname
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
問題: signal.alert_name 存的是 alert_type (如 "custom")，而非 Prometheus
alertname (如 "SentryDown")，導致 playbook Jaccard 匹配永遠失敗 (NO_MATCH)。

根本原因: webhook 的 alertname_to_type mapping 將未知 alertname 轉為 "custom"，
存入 signal.alert_name，但 Playbook 的 symptom_pattern.alert_names 存原始名稱。

修正: 從 signal.labels["alertname"] 讀取原始 Prometheus alertname，
fallback 到 signal.alert_name (保持向下相容)。

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 12:26:18 +08:00
OG T
79a9a514dd fix(rules): ADR-064 L1 Redis 分散式鎖防止多 Pod 重複生成規則
Some checks are pending
CD Pipeline / build-and-deploy (push) Has started running
Details 
問題: _generating set 是進程級，多 Pod 各自獨立，同一 alertname 可能被
  多個 Pod 同時送給 Ollama/Gemini 生成規則

修復: SET NX EX lock_key — 只有第一個 Pod 能取鎖，其他 Pod 直接跳過
降級: Redis 不可用時 fallback 進程級 set（保持原有行為）

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 12:03:51 +08:00
OG T
b66263ad36 fix(decision_manager): resolved Incident 不重送 Telegram
Some checks are pending
CD Pipeline / build-and-deploy (push) Has started running
Details 
dedup TTL 10分鐘過期後，已 resolve 的 Incident 仍被重新推送
加入狀態檢查，resolved/closed 直接跳過

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 12:00:39 +08:00
OG T
9361fd1fa7 fix(decision_manager): action 不應 strip_placeholders 避免截斷 deployment name
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
_strip_placeholders 移除 <...> 導致 kubectl rollout restart deployment/<name>
變成 kubectl rollout restart deployment/，Telegram 顯示建議指令不完整

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 11:45:33 +08:00
OG T
d467fc11be fix(nemotron): 修復 deployment_name placeholder 問題 
根因: Nemotron tool calling 收到 target_resource=DockerContainerUnhealthy
  (非真實 K8s deployment name)，不確定時填 <deployment_name>

修復:
1. prompt 明確標注 deployment_name 必須填入 target_resource
2. 收到 tool call 結果後，偵測 placeholder 並用 target_resource 覆蓋

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 11:44:25 +08:00
OG T
c5e475121a fix(telegram): 修復建議指令被截斷 + decision_manager enum string 補正
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
根因 1: telegram_gateway.py suggested_action[:35] 剛好截到 deployment/ 後
  → 改為 [:80]，完整顯示 kubectl command

根因 2: 舊 Incident proposal_data 存 enum string (RESTART_DEPLOYMENT)
  → decision_manager.py 加入偵測，用規則引擎重新查 kubectl command

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 11:14:30 +08:00
OG T
428e66c111 fix(arch-review): 首席架構師審查 S1×3 S2×3 S3×3 全修復 + ADR-064
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
S1 Critical:
- S1-1: asyncio 觸發移至 _call_with_fallback async 上下文，移除 sync 中的 get_event_loop()
- S1-2: _append_rule_to_yaml 加 textwrap.dedent() 正規化 LLM 輸出縮排
- S1-3: _matches() 對 alertname=["*"] 直接回傳 False，防意外命中

S2 Major:
- S2-1: auto_generate_rule() 改為 DI 參數注入 (ollama_url/model/gemini_api_key)，移除 import settings
- S2-4: _generate_mock_response docstring 澄清為規則引擎生產路徑，非假數據
- S2-5: suggested_action .strip() 防空白字串繞過 or

S3 Minor:
- S3-2: priority 上界 min(next, 890)
- S3-3: alertname sanitize re.sub([{}]) 防 format KeyError
- S3-4: model_registry.py 最後修改時間戳更新

文件:
- ADR-064: Alert Rule Engine YAML 驅動 + AI 自動學習
- Skills 02: 告警規則引擎 DI 規範 + asyncio 禁止事項
- Skills 03: _generate_mock_response 語意澄清 + 規則引擎降級流程
- LOGBOOK: 本次 Session 完整記錄

2026-04-09 ogt: 首席架構師審查修正

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 10:52:40 +08:00
OG T
89da2d24be fix(model-registry): fallback config 更新為 deepseek-r1:14b + gemma3:4b
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 13m20s
Details 
- model_registry._get_default_config: ollama summary llama3.2:3b → gemma3:4b
- model_registry._get_default_config: ollama default/rca → deepseek-r1:14b
- 修正 test_smart_router::test_simple_context 失敗 (斷言 gemma3:4b)
- alert_rule_engine: 移除 asyncio/time unused import
- 2026-04-09 ogt

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 09:52:47 +08:00
OG T
71437db0e9 feat(rule-engine): 自動規則生成 — generic_fallback 觸發 AI 學習
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 11m25s
Details 
流程:
1. 告警命中 generic_fallback 規則
2. 背景觸發 auto_generate_rule()
3. Ollama (deepseek-r1:14b) 生成 YAML 規則片段
4. Ollama 失敗 → Gemini 備援
5. 驗證格式 → append alert_rules.yaml → 清除 lru_cache
6. 下次同類告警直接命中專屬規則，不再走兜底

去重: 同一 alertname 進程內只生成一次
手寫規則 priority 1-499，AI 生成 500-899，兜底 999

2026-04-09 ogt: AI 自學規則引擎

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 09:20:33 +08:00
OG T
d1ede7f989 feat(openclaw): 告警規則引擎 — alert_rules.yaml 取代硬編碼 if/elif
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
- 新增 alert_rules.yaml: 6 條規則 (docker/target_down/oom/cpu/5xx/crash) + 通用兜底
- 新增 alert_rule_engine.py: YAML 載入、匹配邏輯、變數填充
- openclaw.py _generate_mock_response: 重構為呼叫規則引擎 (v8.0)
- 新增規則只需修改 YAML，重啟 Pod 即可，不需改代碼
- 2026-04-09 ogt: 架構重構

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 09:05:23 +08:00
OG T
3abc7c2f85 fix(openclaw): DockerContainerUnhealthy + TargetDown 專屬規則匹配
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
- DockerContainerUnhealthy: ssh docker inspect + docker restart，含 healthcheck 指令驗證
- TargetDown / IP:port instance: ssh 確認 exporter 存活
- 修正 target 混用 alertname 作為 deployment 名稱的問題
- alertname/labels 從 alert_context 提取供規則判斷
- 2026-04-09 ogt: 新增兩條專屬規則

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-09 09:00:31 +08:00
OG T
d80153bdce fix(openclaw): NIM 完全失敗後 fallback 到 Gemini 產生執行方案
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 1m34s
Details 
NIM tool calling 多次 timeout 後，不再顯示空白執行方案，
改由 Gemini 代理產生 kubectl 操作指令（JSON 解析）。
只有 NIM 完全失敗才觸發，符合統帥「必須等到有回應」原則。

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-08 22:55:25 +08:00
OG T
86ac6ed028 perf(api): HostAggregator 效能優化 — probe timeout 縮短 + 30 秒記憶體快取 2026-04-08 22:42:01 +08:00
OG T
c9f1bcd122 fix(api): service_registry 安全降級 — Docker 無 YAML 時不 crash，fallback AUTO
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 11m37s
Details
2026-04-08 21:47:38 +08:00
OG T
db4b28c49d fix(ci): 強制觸發 CD — service_registry.py Docker 路徑修正已包含於 1f9eea5
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 8m45s
Details 
Pod CrashLoopBackOff: IndexError parents[5]
修復: _find_registry_path() 安全搜尋 (parents[4]/parents[3]/絕對路徑)
1f9eea5 已修復但未觸發 CI，此 commit 強制重新 build

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-08 21:37:49 +08:00
OG T
1f9eea5b74 fix(api): service_registry.py Path 索引修正 — 相容 Docker 容器環境
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details
2026-04-08 21:34:40 +08:00
OG T
14cb015826 fix(openclaw): Nemotron 重試邏輯 + exhausted log key (未提交的修改)
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
- generate_incident_proposal_with_tools: 單次 try/except → 2次重試迴圈
- 失敗 log key: nemotron_collaboration_failed → nemotron_collaboration_exhausted
- 失敗時 nemotron_enabled=True (讓統帥看到失敗狀態)
- _call_nemotron_tools: timeout 超時改為拋出異常(讓外層重試)
- 這是之前 Session 的本地修改，修正測試與實際實作不一致問題

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-08 21:16:34 +08:00
OG T
0f5fecfef5 fix(sprint5.1): 首席架構師審查修正 — S1×4 S2×2 S3×1
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 1m40s
Details 
S1-1: service_registry/velero_client/preflight_service 改用 structlog
S1-2: velero_client datetime.now(UTC) 改用 now_taipei()（台北時區鐵律）
S1-3: Guardrail 失敗改為保守拒絕（原放行方向與安全目標相悖）
S1-4: service_registry import 移至模組頂部（移除函數內 import）
S2-1: telegram_gateway T1-T6 六個通知方法補齊 try/except
S2-2: webhooks.py Langfuse URL 改用 settings.LANGFUSE_URL（移除硬寫內網 IP）
S3-3: velero_client trigger_emergency_backup 改為 kubectl apply Backup CRD
      （原 kubectl create backup 語法不存在，審查發現靜默失敗風險）

審查評分: 70/100 → 修正後預計 90+/100

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-08 16:36:18 +08:00
OG T
88696dba9b feat(sprint5.1): Data Safety Guardrails 全鏈路整合 (L1-L5)
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 1m33s
Details
Type Sync Check / check-type-sync (push) Failing after 58s
Details 
Layer 0 - K8s RBAC:
  - k8s/rbac/api-velero-reader.yaml: awoooi-executor SA Velero backup reader

Layer 1 - DB Migration (已在 188 執行):
  - M-002: approval_records 新增 approval_level/votes/required_votes
  - M-003: alert_event_type ENUM 新增 8 個值

Layer 2 - IaC:
  - ops/config/service-registry.yaml: 全服務 Stateful 分級清單 (BLOCK/CRITICAL_HITL/STANDARD_HITL/AUTO)

Layer 3 - Python Services:
  - service_registry.py: 讀取 YAML，提供 is_blocked/requires_multisig/get_required_votes
  - velero_client.py: kubectl 查詢 Velero 備份年齡，失敗 fallback 999h
  - preflight_service.py: Pre-flight 安全檢查 (Q2/Q4 決策)

Layer 1-M001 - Playbook model:
  - playbook.py: 新增 requires_approval_level/stateful_targets/requires_pre_backup

Layer 4 - 業務邏輯:
  - alert_operation_log_repository.py: 新增 8 個 event_type (Guardrail/Pre-flight/MultiSig/備份)
  - auto_repair_service.py: 注入 Service Registry Guardrail 檢查 (BLOCK → 直接拒絕)
  - webhooks.py: ALERT_RECEIVED 溯源記錄 + auto_repair flag Q9 + Langfuse trace_id Q10
  - db/models.py: ApprovalRecord 同步 approval_level/votes/required_votes 欄位
  - docker-health-monitor.sh: 純感知層改造（移除所有 docker restart 邏輯）

Layer 5 - Telegram 通知:
  - telegram_gateway.py: T1-T6 六個新通知方法 (Guardrail/Pre-flight/Backup/MultiSig/ChangeApplied)

參考: ADR-062 Data Safety Guardrails, ADR-063 Service Registry IaC

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-08 16:24:09 +08:00
OG T
eee6f06215 feat(auto-repair): 所有操作強制寫入 DB — auto_repair_executions 表
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 1m32s
Details 
統帥指令: 所有自動修復操作（成功/失敗）必須持久化

變更:
- migrations/phase10_auto_repair_executions.sql: 新增表 + 4 個索引
- db/models.py: 新增 AutoRepairExecution SQLAlchemy model
- repositories/audit_log_repository.py: 新增 AutoRepairExecutionRepository (create/list_by_incident/get_stats)
- auto_repair_service.py: execute_auto_repair 成功/失敗分支都寫入 DB
  - 新增 similarity_score 參數傳遞
  - AutoRepairDecision 新增 similarity_score 欄位
- webhooks.py: 傳入 similarity_score 到 execute_auto_repair

已執行 migration: awoooi_prod@192.168.0.188:5432 ✅

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
 2026-04-08 11:16:37 +08:00
OG T
68a2fff746 feat(auto-repair): 移除所有阻擋門檻 — 直接全部跳成自動修復
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 1m38s
Details 
統帥指令: 所有 APPROVED Playbook 直接執行，不再檢查:
- 相似度門檻 (MIN_SIMILARITY_SCORE 0.7 → 0.0)
- is_high_quality 品質門檻
- 冷啟動信任機制
- 動作風險等級門檻 (evaluate + execute 兩層)

保留: P0/P1 嚴重度人工審核、全域冷卻熔斷、APPROVED 狀態檢查

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
 2026-04-08 11:10:09 +08:00
OG T
b7ea362efc fix(api): Review #2 技術債清理 — I1/S1/S2/S3 全數修正
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 12m13s
Details 
I1: error_type 欄位補全
- AnomalyCounter.derive_key_from_incident() 新增
  從 signal.labels 提取 reason/error_type，確保四欄位完整

S1: 三處 signature 建構邏輯統一
- auto_repair_service._derive_anomaly_key() → 委託 derive_key_from_incident()
- approval_execution._get_anomaly_key_from_approval() → 同上
- incident_service.resolve_incident() B4 → 同上
- 消除 3 處重複的 signature 建構程式碼

S2: Redis Pipeline 批次查詢
- get_all_disposition_stats() 從 N+1 hgetall 改為 2 次 Pipeline
- Pipeline 1: 批次 hgetall 所有 disposition key
- Pipeline 2: 批次 hget metadata (alert_name)
- 效能從 O(2N) Redis round-trip 降至 O(2)

S3: auto_repair.py get_incident AttributeError 修復
- get_incident() → get_from_working_memory() (pre-existing bug)

Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
 2026-04-07 13:13:42 +08:00
OG T
de3935d1d4 feat: Sprint 4 Phase E+F — 前端處置統計 + 週報處置分佈
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 1m26s
Details
Type Sync Check / check-type-sync (push) Failing after 1m2s
Details 
Phase E: 前端頁面
- E1: /reports 完整處置統計儀表板 (已在 Sprint F 完成)
- E2: 首頁 Metrics Strip — 從 disposition API 取得真實自動化率
  優先使用 /stats/disposition auto_rate，fallback 到 incidents 推算
- E3: /auto-repair 處置概況卡片 (已在 Sprint F 完成)
- E4: /neural-command stats tab 處置分佈 (已在 Sprint F 完成)
- E5: i18n 翻譯 zh-TW + en (已在 Sprint F 完成)

Phase F: 週報 + 文件
- F1: WeeklyReportMessage 新增 disposition 5 欄位
  週報格式加「📋 處置分佈」區塊 (自動/冷啟動/人工/手動 + 自動化率)
  weekly_report_service 整合 get_all_disposition_stats()
- message 字數上限從 900 提升到 1200 (適應處置區塊)

Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
 2026-04-07 13:02:20 +08:00
OG T
561bcb638b fix(api): Sprint 4 首席架構師 Review P0 修正 — hash 統一 + 積木化合規 
P0-1: anomaly_key hash 推導統一
- B1: 新增 _derive_anomaly_key() 使用 AnomalyCounter.hash_signature()
  取代 symptoms.compute_hash()
- B3/B4: namespace 改用 signal.labels.get("namespace", "")
  修正 getattr(signal, "namespace", "") 永遠回傳空字串

P0-2: Router 層積木化合規
- C1/C2: 封裝 get_all_disposition_stats() 到 AnomalyCounter
- Router 不再直接存取 counter.redis
- stats.py 移除未使用的 days/stats 參數

P1: get_frequency() 填充 disposition 欄位
- 與 _record_anomaly_impl() 一致，回傳完整處置統計

首席架構師評分: 82/100 → P0 全數修正

Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
 2026-04-07 12:53:12 +08:00
OG T
a85e9ced08 feat(api+telegram): Sprint 4 Phase C+D — API 端點 + Telegram 處置統計 
Phase C: API 端點
- C1: GET /api/v1/stats/disposition — 完整處置分佈統計
  - DispositionSummary: auto/human/manual/cold_start + auto_rate
  - DispositionByAnomaly: 按異常類型明細 (最多 20 筆)
  - Redis SCAN + HGETALL 聚合
- C2: GET /api/v1/auto-repair/stats 擴充 disposition_summary

Phase D: Telegram 告警格式
- D1: 告警卡片加處置統計行
  - 🤖 自動: N | 👤 審核: N | 🔧 手動: N
  - 自動化率百分比
- D2: 歷史按鈕強化處置分佈明細
  - 完整 5 項計數 + 自動化率

Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
 2026-04-07 12:17:20 +08:00
OG T
9253281d46 feat(api): Sprint 4 Phase A+B — 告警處置統計資料層+寫入層 
Phase A: 資料層
- A1: IncidentFrequencyStats 新增 4 欄位 (human_approved/manual_resolved/cold_start_trust/total_resolution)
- A2: AnomalyCounter.record_disposition() — Redis HINCRBY 原子遞增
- A3: get_disposition_stats() — HGETALL 回傳處置分佈
- AnomalyFrequency dataclass 擴充 + to_dict() 同步
- _record_anomaly_impl() 整合 disposition stats

Phase B: 寫入層觸發點接線
- B1: 自動修復成功 → record_disposition("auto_repair")
- B2: 冷啟動信任成功 → record_disposition("cold_start_trust")
  - AutoRepairDecision 新增 is_cold_start flag
  - execute_auto_repair() 接收並區分處置類型
- B3: 人工批准執行成功 → record_disposition("human_approved")
  - 新增 _get_anomaly_key_from_approval() helper
- B4: 手動處理推斷 → resolve_incident() 排除法判定
  - 若 resolved 且無 auto/human/cold_start 紀錄 → manual_resolved

安全設計: 所有 disposition 記錄走 try/except，失敗不阻塞主流程

Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
 2026-04-07 11:54:46 +08:00
OG T
53b2daeaca feat(api): 首次信任機制 — 打破自動修復冷啟動雞生蛋問題 
問題: Playbook 需要 success_count >= 3 才算 is_high_quality，
但沒有自動修復就不會有成功紀錄 → 永遠達不到門檻。

方案 C: 首次信任 (Cold Start Trust)
- APPROVED 狀態 + 全步驟 risk=LOW + 執行次數 < 3 → 自動放行
- Redis counter 限制每日最多 5 次首次信任自動修復
- 累積 3 次成功後自動回歸正常 is_high_quality 門檻

安全邊界:
- 只有 LOW risk 步驟才能首次信任 (重啟容器等)
- HIGH/CRITICAL 仍需人工審核
- P0/P1 嚴重度仍需人工審核
- 每日上限防止失控

Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
 2026-04-07 11:21:00 +08:00
OG T
2fe8062fb8 refactor(api): Re-Review S1/S2/S3 改善 — 消除重複+防禦性驗證+測試隔離 
S1: 抽取 _execute_and_observe() 公用方法
  - 消除 repair_by_uri 中 3 處重複的 execute+audit+langfuse 邏輯
  - 統一 AuditLog + Langfuse trace 寫入路徑

S2: SSH username 防禦性驗證
  - 新增 validate_ssh_user() + _SSH_USER_RE 正則
  - 在 _ssh_execute() 入口驗證 user 參數
  - 防止 user@host 拼接產生非預期行為
  - 新增 8 個 username 驗證測試

S3: Singleton 測試重置
  - 新增 _reset_for_test() classmethod
  - 避免跨測試狀態污染
  - 新增 2 個 singleton reset 測試

測試: 55/55 全數通過 (原 45 + 新 10)
首席架構師 Re-Review: 91/100 ✅ 通過，3 個 Suggestion 全數實裝

Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
 2026-04-07 11:17:40 +08:00
OG T
78a8d3dfa5 fix(api): ansible 控制節點加白名單驗證，防環境變數繞過 (Re-Review Important) 
首席架構師 Re-Review 指出: ANSIBLE_CONTROL_HOST 來自環境變數 (ConfigMap)，
若 ConfigMap 被篡改可繞過 SSH_TARGET_WHITELIST。
在 _execute_ansible() 開頭加 validate_ssh_target_host(host) 閉環。

Re-Review 評分: 91/100 ✅ 通過

Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
 2026-04-07 11:13:49 +08:00
OG T
f8d4772abf fix(api): Sprint 3 P0-1/P0-2/P0-3/P0-4 Critical Security Fixes 
P0-1: Complete shell metacharacter regex detection
  - Enhanced _SHELL_METACHAR_RE to detect: >, <, \n, ${}, $()
  - Prevents all shell injection vectors (redirects, variable expansion, newlines)
  - Added 5 new validation tests

P0-2: Add shlex.quote() protection for ansible playbook path
  - Wraps playbook_path in shlex.quote() before SSH command construction
  - Prevents shell injection if path contains special characters
  - Applied in _execute_ansible() method

P0-3: Add SSH target host whitelist validation
  - Introduces validate_ssh_target_host() function
  - Only allows SSH to: 192.168.0.110, 192.168.0.188
  - Prevents unauthorized SSH target exploitation
  - Added 5 new whitelist validation tests

P0-4: Convert HostRepairAgent to singleton pattern
  - Implements __new__() singleton with shared _in_process_locks dict
  - Ensures in-process locks persist across multiple auto_repair_service calls
  - Previously created new instance per call, making locks ineffective
  - Added singleton persistence test

Test Results: 45/45 passing (34 existing + 11 new P0 tests)
All security validations verified via comprehensive unit test coverage.

Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
 2026-04-07 11:09:45 +08:00
OG T
af07c23675 fix(k8s): known_hosts 改掛 /etc/repair-known-hosts 獨立目錄，修 mount 衝突
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 12m11s
Details
E2E Health Check / e2e-health (push) Successful in 34s
Details 
/etc/repair-ssh 已被 repair-ssh-key 佔用，subPath 檔案掛載衝突
改為獨立目錄 /etc/repair-known-hosts，路徑同步更新 KNOWN_HOSTS_PATH

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-06 15:06:28 +08:00