wooo/awoooi
1
1
Fork 0
Code Issues 2 Pull Requests 1 Actions 3 Packages Projects Releases Wiki Activity
1,184 Commits 179 Branches 2 Tags
a2cc985f601f4dcf9fc6286ed3f4792ef710ca76
Commit Graph

545 Commits

Author SHA1 Message Date
OG T
a2cc985f60 feat(mcp-phase3): ArgoCD MCP + Sentry MCP + 完整 Provider 註冊
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
ArgoCDProvider (3 工具):
  - argocd_list_apps: 列出所有 App + sync/health 狀態
  - argocd_get_app_status: 詳細狀態 + 問題資源清單
  - argocd_get_sync_history: 最近 N 筆部署記錄
  - 輸入驗證: app_name 白名單 regex
  - 需 ARGOCD_API_TOKEN + ARGOCD_MCP_ENABLED=true

SentryProvider (3 工具):
  - sentry_list_issues: 列出最近 Issues（狀態過濾）
  - sentry_get_issue: 詳情 + stacktrace 最後 5 frames
  - sentry_search_issues: PromQL 風格搜尋
  - issue_id 白名單驗證（只允許純數字）
  - 需 SENTRY_AUTH_TOKEN + SENTRY_MCP_ENABLED=true

providers/__init__.py: 補上 Prometheus + SSH + ArgoCD + Sentry 全部 10 個 providers
config.py: 新增 ARGOCD_URL / ARGOCD_API_TOKEN / ARGOCD_MCP_ENABLED / SENTRY_MCP_ENABLED

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-11 09:11:53 +08:00
OG T
1ec19656b5 feat(adr071-ij): TYPE-2 指標快照卡片 + KM 三段資料整合
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 8m17s
Details
Deploy Alert Rules / Deploy Prometheus Alert Rules (push) Successful in 36s
Details
Ansible Lint / lint (push) Has been cancelled
Details 
ADR-071-I: decision_manager 執行前後各抓一次 Prometheus metrics
  - _fetch_metrics_snapshot(): 依 alertname 選擇 CPU/Mem/Disk/Restart 查詢
  - _format_metrics_delta(): 輸出 "CPU 92%→23% | Mem 78%→45%" 格式
  - _push_auto_repair_result(): metrics_after 寫 DB + TYPE-2 卡片顯示 delta
  - _auto_execute(): metrics_before 在執行前寫 DB（完成閉環）

ADR-071-J: km_conversion_service._build_content() 使用精簡 delta 格式
  - 從 metrics_before/after 產生人讀 delta（CPU/Mem/Disk/重啟次數）
  - 附加 k8s_state_after（若有）
  - 格式: 症狀 + 根因 + 動作 + 效果數字（症狀→情境→動作→效果）

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-11 03:09:35 +08:00
OG T
a29e5e1de2 feat(mcp-phase1): K8s MCP 強化 — 6 個新工具 + namespace 白名單 
MCP Phase 1 (ADR-069 Sprint B 後驗收):
  k8s_get_pod_logs    — Pod log 取得 (tail 1-500，支援 previous)
  k8s_watch_rollout   — rollout 狀態監控直到完成 (timeout 10-300s)
  k8s_get_events      — K8s events (可過濾 resource_name / event_type)
  k8s_describe_pod    — 完整 Pod describe (Conditions/Volumes/Env)
  k8s_get_hpa_status  — HPA 副本數/CPU utilization
  k8s_get_node_conditions — Node Ready/MemoryPressure/DiskPressure

安全強化:
  - ALLOWED_NAMESPACES = {"awoooi-prod"} 硬編碼白名單
  - _validate_namespace() + _validate_name() 參數白名單
  - 數值參數上下限夾緊 (tail 1-500, timeout 10-300s)
  - event_type 只允許 Warning / Normal

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-11 03:01:38 +08:00
OG T
2af4dffcc6 fix(security): Architecture Review 修復 5 項高信心問題 
安全修復 (P0):
1. ssh_provider: 新增 _validate_param() 白名單驗證，防止 command injection
   - container_name/service/filter_name: [a-zA-Z0-9._-]{1,128}
   - compose_dir: 必須以 /opt/ 或 /srv/ 開頭，禁止 ..
   - domain: FQDN 白名單
   - tail/port/lines: int() 轉換 + 上下限夾緊
2. ssh_provider: known_hosts=None 改為讀 SSH_MCP_KNOWN_HOSTS_FILE 環境變數
   - 預設仍 None（內網快速啟動），但啟動時寫入 warning log
   - 設定文件：ops/runbooks/ssh-mcp-setup.md (待補)

模組化修復 (P1):
3. km_conversion_service: 移除 import 時的 ALERT_EVENT_TYPES.update() 副作用
   - ADR-071 event types 移入 alert_operation_log_repository.py 靜態集合
4. telegram_gateway: create_task() 改為 await + try/except
   - 避免 DB session 關閉後的競爭條件
   - KM 轉換失敗記錄 warning log，不中斷主流程
5. km_conversion_service: 新增頂層 try/except，錯誤一律 error log 後 re-raise

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-11 02:50:26 +08:00
OG T
6351e9a0e9 feat(mcp-phase2): MCP Phase 2 — Prometheus MCP + SSH MCP + alert labels
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 13m37s
Details
Deploy Alert Rules / Deploy Prometheus Alert Rules (push) Successful in 35s
Details 
MCP-2b: prometheus_provider.py
  - prometheus_query (PromQL 即時查詢)
  - prometheus_query_range (歷史趨勢，預設 15 分鐘)
  - prometheus_get_alert_history (告警觸發歷史)
  - config: PROMETHEUS_URL + PROMETHEUS_MCP_ENABLED

MCP-2a: ssh_provider.py
  - 群組A 9 個只讀診斷工具 (top/disk/memory/logs/status/port/nginx/swap)
  - 群組B 6 個安全操作工具 (restart/compose/systemctl/clear-log/ssl/nginx-reload)
  - 四層安全守衛 (白名單/allowed_hosts/forbidden_patterns/trust_score)
  - config: SSH_MCP_ENABLED + SSH_MCP_ALLOWED_HOSTS

K8s: 04-ssh-mcp-secret.example.yaml (ssh-mcp-key Secret 範本 + 建立步驟)

Alert labels: alerts-unified.yml 補充 mcp_provider/host_type/alert_category
  覆蓋: HostHighCpuLoad/HostOutOfMemory/HostOutOfDiskSpace/DockerContainer*
        SignOzDown/SentryDown/HarborDown/GiteaDown

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-11 02:35:35 +08:00
OG T
325b3851b5 feat(adr-071): 告警通知四類型第一批 B/C/E/F/G/H 全實作
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details
Type Sync Check / check-type-sync (push) Failing after 1m7s
Details 
ADR-071-B: classify_notification() — 五型分類器 (TYPE-1/2/3/4/4D)
ADR-071-C: send_info_notification() — TYPE-1 純資訊無按鈕卡片
ADR-071-E: _build_inline_keyboard() — 依 alert_category 動態組合 TYPE-3 按鈕
ADR-071-F: send_drift_card() — TYPE-4D Config Drift 卡片 + Diff 截斷
ADR-071-G: km_conversion_service.py — Incident RESOLVED 自動轉 KM
ADR-071-H: handle_manual_fix_done() — TYPE-4 手動修復 Bot 對話閉環

前批已完成: ADR-071-A (DB Migration) + ADR-071-D (狀態機守衛)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-11 02:24:20 +08:00
OG T
68a3858ae4 fix(auto_execute): 守衛加入 target==alertname 檢查，防止 LLM 把告警名稱當 deployment 名稱
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 13m33s
Details 
HostHighCpuLoad 等主機告警，NemoTron Tool Calling 可能把
alertname 填入 deployment_name，導致執行
'kubectl rollout restart deployment HostHighCpuLoad'。

新增守衛: _target == _alertname 時拒絕執行並通知人工介入。

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-11 01:13:24 +08:00
OG T
a4d655ea7f fix(auto_execute): 安全守衛 — 拒絕執行含 unknown 或未解析 placeholder 的 action
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 19m7s
Details
E2E Health Check / e2e-health (push) Successful in 43s
Details 
主機層告警（HostHighCpuLoad、DockerContainerUnhealthy 等）沒有對應
K8s deployment 名稱，affected_services=[]，導致 _target='unknown'，
執行 'kubectl rollout restart deployment unknown' 這種無意義命令。

修復: 替換後若 action 仍含 'unknown' 或 <...>/{...} 格式，
直接拒絕執行並通知人工介入，不允許帶 placeholder 的命令上線。

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 23:57:17 +08:00
OG T
dabc62e0f8 fix(telegram): append_incident_update — 儲存告警卡片 message_id 到 Redis
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 14m31s
Details 
_send_approval_card_to_group 發出告警卡片後，將 Telegram message_id
存入 Redis tg_msg:{incident_id}（TTL 24h），供後續 append_incident_update
換掉批准按鈕 + reply 狀態。

修復前：tg_msg key 從未被寫入，append 永遠 fallback 發新訊息，
批准按鈕永遠無法被移除。

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 22:41:30 +08:00
OG T
797c7c749e fix(nemotron): deepseek-r1 num_predict 400→1200，避免 <think> block 截斷後空回覆
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 28s
Details 
deepseek-r1:14b 思考 token 超過 400 會在 </think> 前截斷，導致
清理後 body 為空，Telegram 顯示空訊息。
- chat_manager: num_predict 400 → 1200
- telegram_gateway: _clean_ai_reply 空值加 fallback 錯誤提示

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 22:35:37 +08:00
OG T
d1c85c332a feat(models): models.json v1.3 — 加入 ADR-067 五大 Ollama 應用設定
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 14m21s
Details 
新增 adr067_ollama_applications 區塊:
- Phase 30: drift_summary (qwen2.5:7b-instruct, 90s)
- Phase 31: log_anomaly_summary (deepseek-r1:14b, 120s)
- Phase 32: pr_code_review (qwen2.5-coder:7b, 120s)
- Phase 33: rag_embed (nomic-embed-text 768d) + rag_generate (qwen2.5:7b)
- Phase 34: image_analysis (llava:latest, 60s)

endpoint 統一標注 http://192.168.0.111:11434 (ADR-067 專屬)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 22:16:09 +08:00
OG T
f8926bb70a ci: 觸發 CD — decision_manager 修復標記 
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 22:12:56 +08:00
OG T
e5f1541d69 fix(auto_execute): 替換 action 中的 <deployment_name>/{target}/{namespace} placeholder
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 24s
Details 
Nemotron tool calling 生成 <deployment_name> 佔位符未替換
auto_execute 前統一替換所有 {target}/{namespace}/<xxx> 格式

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 22:00:19 +08:00
OG T
71f0dbf2b5 fix(auto_execute): ApprovalRequest 補齊 description/requested_by/required_signatures
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
3 validation errors 導致 auto_execute_failed
補上所有必填欄位，required_signatures=0 表示自動核准不需簽核

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 21:59:52 +08:00
OG T
f33d514391 fix(auto_repair): playbook_seed_service — 從 alert_rules.yaml 初始化 APPROVED Playbook
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
根本原因: playbooks 表空 → NO_MATCH → 永遠走審批，從不自動修復
修復: startup 時從 alert_rules.yaml seed APPROVED Playbook（冪等）
確保自動修復鏈路有規則可用

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 21:52:38 +08:00
OG T
100e4d9b89 fix(chat): AI 回覆截斷問題 — 強制 persona + Markdown 清理 + 600字上限
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 14m39s
Details 
問題: OpenClaw/NemoClaw 回覆 Markdown 語法 + 超長，Telegram 顯示截斷
修正:
1. chat_manager: _call_openclaw/_call_nemotron 強制前置 persona (含不超過300字規範)
2. telegram_gateway: _clean_ai_reply() 移除 **bold** *italic* # header 語法
   移除 deepseek-r1 <think> 標籤，截斷 > 600 字並在段落邊界截

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 21:26:15 +08:00
OG T
527ce9faaf fix(notifications): 新增後端 /api/v1/notifications/channels 路由
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 2m4s
Details 
前端 /notifications 頁面呼叫此 endpoint 但後端不存在 (404)
新增 notifications.py：回傳 4 個真實頻道狀態
- Telegram OpenClaw Bot (BOT_TOKEN 設定檢查)
- Telegram Nemotron Bot (BOT_TOKEN 設定檢查)
- SSE Web Stream (永遠 active)
- Redis Stream awoooi:signals (ping 檢查)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 16:17:37 +08:00
OG T
167e115a6d feat(phase31): Log 異常摘要觸發點 — 告警後 NemoTron 發 log summary
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 2m44s
Details 
_send_log_summary: 取 Pod log → deepseek-r1:14b 分析 → NemoTron 發到群組
觸發點: _push_decision_to_telegram 送完審批卡後異步執行

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 16:07:56 +08:00
OG T
95f63d64d7 fix(auto_approve): min_trust_score 0 解除自動修復封鎖
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
根本原因: trust_score 是 in-memory dict，Pod 重啟即歸零
永遠 < min_trust_score=1 → 所有告警走審批，從未自動執行

修復: min_trust_score=0，medium risk + confidence>=0.65 直接自動執行

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 16:06:40 +08:00
OG T
ff3be51e13 fix(phase34): 圖片分析改用 send_as_openclaw 發到 SRE 群組
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
send_notification 發到私人 chat，改用 send_as_openclaw 發到 SRE 戰情室

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 15:56:19 +08:00
OG T
b9dbbb3575 feat(rag): Telegram /rag 指令 + /rag/optimize ivfflat 端點
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 14m9s
Details 
- telegram_gateway: /rag <query> → KnowledgeRAGService.query()
  _handle_group_command 加 full_text 參數取得完整指令文字
  /help 更新加入 /rag 說明
- rag.py: POST /rag/optimize → rag_repo.create_ivfflat_index()
- rag_chunk_repository: create_ivfflat_index() — ivfflat with lists=100

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 14:47:21 +08:00
OG T
33abe988f8 fix(phase34): 圖片分析結果改由 OpenClaw 回覆（llava vision）
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
NemoTron 負責文字問答(deepseek-r1:14b)，OpenClaw 負責圖片分析(llava)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 14:13:57 +08:00
OG T
7e5ac00d62 fix(phase34): image_analysis 用正確 bot token 下載 + NemoTron 回覆
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
- 下載圖片改用 OPENCLAW_TG_BOT_TOKEN（polling bot）
- 結果改用 send_as_nemotron 從 NemoTron bot 回覆到群組

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 13:58:59 +08:00
OG T
cf5eb71ea6 fix(phase34): polling loop 補圖片路由 — _handle_chat_message photo handler
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
text=None 時直接 return，導致圖片訊息被丟棄
在 text 檢查前插入 photo 路由，呼叫 image_analysis_service

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 13:58:05 +08:00
OG T
e65d931e73 fix(ci): B5 整合測試 DinD 修正 — 用 host bridge IP + published port
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 2m2s
Details 
DinD 環境下 volume mount 和 compose 網路都不可靠：
- runner container 的路徑 ≠ host 路徑 (volume 失敗)
- compose 網路 IP 對 runner 不可路由

解法: host docker bridge (ip route default gateway) + postgres exposed port 15432
runner 直接用 /opt/api-venv/bin/pytest (host runner 上已安裝)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 13:03:25 +08:00
OG T
c8b5c994d4 fix(ci): B5 整合測試改用 compose pytest-runner service
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 2m27s
Details 
問題: runner 是 Docker-in-Docker，-v /opt/api-venv 掛到 host 路徑不存在

修正:
- docker-compose.test.yml: 新增 pytest-runner service (python:3.11-slim)
  在 compose 網路內跑，hostname=postgres-test 直連，自裝 deps
- postgres-test: initdb.d 自動執行 setup_test_schema.sql，省去手動 psql
- cd.yaml: 改用 --profile test + --exit-code-from pytest-runner

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 12:58:44 +08:00
OG T
7768924fea fix(flywheel): 自動修復後移除 Telegram 按鈕 + 心跳告警排除飛輪
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 6m56s
Details 
問題: 自動修復成功後 Telegram 卡片仍顯示批准/拒絕/靜默按鈕

Fix 1 — Telegram 卡片回饋閉環 (積木化合規):
- telegram_gateway.send_approval_card: 發送後自動存 tg_approval:{id} 到 Redis
- telegram_gateway.mark_auto_repaired(): 新方法 — 移除按鈕 + reply 結果
- _try_auto_repair_background: 改呼叫 gateway.mark_auto_repaired() (Service 層)

Fix 2 — 心跳/看門狗告警排除飛輪:
- constants.py: is_heartbeat_alertname() + HEARTBEAT_ALERT_NAMES
- NoAlertsReceived2Hours 等不觸發 _try_auto_repair_background

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 11:52:04 +08:00
OG T
485b8cb003 fix(ci): B5 整合測試加 ssl=disable — asyncpg 預設嘗試 SSL 被 container 拒絕
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 1m55s
Details 
錯誤: ConnectionRefusedError Connect call failed ('127.0.0.1', 15432)
根因: asyncpg 走 _create_ssl_connection，臨時 postgres container 無 SSL
修正: TEST_DATABASE_URL + conftest 預設 URL 均加 ?ssl=disable

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 11:40:40 +08:00
OG T
670cd5df86 refactor(flywheel): 首席架構師審查修正 C1/C2/I1/I2/I3/I4/M1
Some checks are pending
CD Pipeline / build-and-deploy (push) Has started running
Details 
C1 — Repository 層修正 (積木化鐵律):
  新增 PlaybookEmbeddingRepository (pgvector UPSERT)
  playbook_embedding_service 改透過 Repository 存取 DB，不再直接 db.execute(text(...))

C2 — Router 層業務邏輯移入 Service 層:
  create_incident_for_approval + extract_affected_services (去掉底線前綴) 移入 incident_service.py
  webhooks.py 改從 incident_service import，自身不再含業務邏輯

I1 — _infra_jobs 提升為 module-level frozenset (_INFRA_JOB_NAMES)，避免每次呼叫重建

I2 — _persist_embeddings_to_db 補齊 PlaybookRAGService / list[Playbook] 型別標注

I3 — embedding 格式顯式化: "[" + ",".join(str(float(x)) for x in embedding) + "]"
  防止 pgvector 因格式差異靜默解析失敗

I4 — import asyncio 移至 main.py 頂層，移除 try 區塊內重複 import

M1 — similarity.py: 移除死代碼 `if union > 0 else 0.0`
  union 在兩個集合都非空時不可能為 0

2026-04-10 Asia/Taipei — Claude Sonnet 4.6
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 11:35:10 +08:00
OG T
49bfbd573c feat(test): B5 整合測試框架 — 真實 DB, 5/5 通過
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 2m34s
Details 
新增:
- docker-compose.test.yml: CI 用臨時 pgvector PostgreSQL (port 15432)
- tests/factories.py: Incident/Approval/Knowledge/RAG 測試資料工廠
- tests/integration/test_b5_core_flows.py: 5 個 E2E 整合測試 (5/5 PASSED 1.03s)
- tests/integration/setup_test_schema.sql: CI schema 初始化 SQL
- cd.yaml: 新增 Integration Tests B5 step
- scripts/sync_dev_db.py: dev DB 同步工具

修正:
- .env.test: DATABASE_URL 指向 awoooi_dev (本機設定, gitignore 不入庫)

禁止 Mock 鐵律: 所有 DB 測試使用真實 PostgreSQL, 無 SQLite/MagicMock

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 11:22:57 +08:00
OG T
ab6f6faa32 feat(phase32): 實作 review_push + gitea_webhook 改用本地 Ollama 審查
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
- local_code_review_service: 新增 review_push() 方法
  使用 qwen2.5-coder:7b 審查 push event（非 PR）
- gitea_webhook_service: _call_openclaw_push_review 改用本地推理
  OpenClaw 無 push-review 端點(404) → 改用 LocalCodeReviewService

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 11:09:11 +08:00
OG T
b24fae313e fix(drift_narrator): 補寫 narrative_text 到 DB + drift_repository.update_narrative
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details
2026-04-10 11:06:50 +08:00
OG T
c6edfb5614 fix(flywheel): 四階段系統性修復 AUTO_REPAIR NO_MATCH 斷層
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
Phase 1 — affected_services 污染根治
  - webhooks.py: _extract_affected_services() 從 labels 精準萃取服務名
    (component > job > pod deployment name > clean target_resource > [])
  - create_incident_for_approval: alert_labels 完整保留進 Signal
  - alert_name 從 alertname 取，不再用 "custom"

Phase 2 — Playbook alertname 變體擴充
  - alert_rules.yaml: 5 條規則新增 HostHighCpuLoad、KubePodCrashLooping 等變體
  - scripts/update_playbook_alert_variants.py: Redis index 已執行更新 ✅

Phase 3 — Jaccard 通用型 Playbook 豁免
  - similarity.py: affected_services=[] → 1.0 豁免（基礎設施 Playbook 不針對特定服務）
  - severity_range=[] → 1.0 豁免（適用所有嚴重度）

Phase 4 — Playbook Embedding 持久化（冷啟動修復）
  - migrations/flywheel_playbook_embeddings.sql: pgvector 持久化表
  - services/playbook_embedding_service.py: 啟動時重建 Redis 向量快取 + 同步 DB
  - main.py: lifespan 啟動時 asyncio.create_task 非阻塞執行

2026-04-10 Asia/Taipei — Claude Sonnet 4.6
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 11:04:56 +08:00
OG T
1c4bdedc64 fix(drift_narrator): send_text → send_notification + DriftLevel case fix
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 14m43s
Details
2026-04-10 10:48:36 +08:00
OG T
5c2db65ea1 refactor(rag): C1 修正 — 新增 rag_chunk_repository，Service 不再直接存 DB
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
- 新增 src/repositories/rag_chunk_repository.py
  search_chunks / insert_chunk / delete_by_source_id / get_stats
- KnowledgeRAGService 移除所有 get_db_context 直接呼叫
  改委派 rag_repo.search_chunks / insert_chunk / delete_by_source_id / get_stats
- 移除 unused Any import

leWOOOgo 合規評分: 62 → 95/100

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 10:43:53 +08:00
OG T
cc8cabebf9 refactor(rag): 架構審查修正 — leWOOOgo 合規 + 去重 + httpx 關機
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
- C2: _run_index() 業務邏輯移入 KnowledgeRAGService.index_all_sources()
      Router 層只做 background_tasks.add_task(_run_index) 轉發
- C3: glob("*.md") → rglob("*.md") — 掃描巢狀子目錄
- C4: docstring 修正 Ollama 188 → 111
- I2: index_document() 先刪舊版本 (_delete_by_source_id) 避免重複累積
- I3: debug endpoint 改用 settings.OLLAMA_URL 取代硬碼 IP
- I4: main.py shutdown 加入 get_knowledge_rag_service().close()

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 10:39:14 +08:00
OG T
af7b1591c1 feat(rag): phase35 ivfflat 向量索引 — 5814 chunks 已建立
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
已在 prod 執行: idx_rag_chunks_embedding (lists=100, cosine)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 10:33:32 +08:00
OG T
09a8c3a90b fix(rag): 修正 debug endpoint 與訊息文字 — Ollama 188→111
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 10:28:04 +08:00
OG T
68e9ef5d26 fix(drift_narrator): DriftItem.severity → drift_level.value 欄位名稱修正
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details
2026-04-10 10:24:41 +08:00
OG T
974f84511b fix(rag): embed 改用 settings.OLLAMA_URL — K3s NetworkPolicy 擋住直連 188:11434
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
nomic-embed-text 在 111 也有，改走 OLLAMA_URL (111) 避開 NetworkPolicy

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 10:14:33 +08:00
OG T
b51f1b011c debug(rag): /rag/debug 顯示完整 Ollama 錯誤訊息
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 10:13:52 +08:00
OG T
6786da89c8 debug(rag): 加入 /rag/debug 診斷端點 — 確認容器路徑 + Ollama 連線
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 13m14s
Details 
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 09:54:56 +08:00
OG T
b261a51685 feat(rag): Dockerfile 加入 docs/ + .agents/skills/ — RAG 索引來源
Some checks failed
CD Pipeline / build-and-deploy (push) Failing after 2m11s
Details 
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 09:16:51 +08:00
OG T
3ed52b0424 fix(rag): _run_index 修正 index_document 簽名不符 — 讀檔內容再傳 service
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 13m3s
Details 
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 09:00:26 +08:00
OG T
0ee5d532ba feat(rag): 新增 RAG Router + 掛載到 main.py (Phase 33 ADR-067)
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 13m11s
Details 
- rag.py: POST /index, POST /query, GET /stats 三端點
- stats 委派給 KnowledgeRAGService.get_stats()（leWOOOgo 合規）
- main.py: include_router rag_v1.router

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 07:34:06 +08:00
OG T
e605b7192b feat(rag): Phase 33 RAG API 端點 — /knowledge/rag/index + query + stats
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 14m35s
Details 
ADR-067 Phase 33: pgvector RAG 三個 HTTP 端點
- POST /knowledge/rag/index — 索引文件到 rag_chunks
- GET /knowledge/rag/query — embed→knn→生成答案
- GET /knowledge/rag/stats — chunks 統計 (透過 Service 層)
- 修正: rag_stats 移至 KnowledgeRAGService.get_stats() (leWOOOgo 積木化)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 02:00:59 +08:00
OG T
63e840ae42 feat(ollama): Phase 31-34 ADR-067 — Log摘要/PR審查/RAG知識庫/圖片分析
Some checks are pending
CD Pipeline / build-and-deploy (push) Has started running
Details 
Phase 31: log_summary_service.py — deepseek-r1:14b K8s Pod日誌異常摘要
  - 觸發: signoz_webhook 告警時背景呼叫
  - Redis快取 log_summary:{pod}:{date} TTL 24h
  - 敏感資料regex遮蔽

Phase 32: local_code_review_service.py — qwen2.5-coder:7b PR自動審查
  - Fallback: Gemini (diff > 50KB 或 Ollama超時)
  - semaphore 最多2個同時審查
  - 雙寫: Redis TTL 7d + pr_reviews表 (phase29 migration)

Phase 33: knowledge_rag_service.py — nomic-embed-text 768維 pgvector RAG
  - 向量化(188) + 生成(111) 雙Ollama
  - rag_chunks表 (phase28 migration)
  - 初期線性搜尋，>100筆啟用ivfflat索引

Phase 34: image_analysis_service.py — llava:latest Telegram圖片分析
  - download_and_analyze: Bot API getFile → 下載 → llava → 回應
  - Rate limit: 每chat_id每分鐘3次 (Redis sliding window)
  - telegram.py webhook新增photo分支

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 01:50:22 +08:00
OG T
89015d4527 feat(phase30): Drift 報告 AI 人話摘要 (ADR-067)
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
- 新增 DriftNarratorService — qwen2.5:7b-instruct (Ollama 111)
  - 觸發條件: high >= 1 or medium >= 3（HPA replicas 白名單）
  - Redis 快取: drift_narrative:{report_id} TTL 1h
  - LLM 失敗時 graceful fallback 結構化文字
- drift.py _analyze_and_notify: 接入 narrator（Phase 30 標記）
- Migration: drift_reports.narrative_text TEXT (已在 prod 執行)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 01:37:43 +08:00
OG T
a30713b292 fix(chat): NemoClaw 禁止自稱 DeepSeek + 強制繁體中文
All checks were successful
CD Pipeline / build-and-deploy (push) Successful in 13m36s
Details 
- 明確禁止透露底層模型身分
- 強制繁體中文（禁簡體）
- 補充 SRE 專長範圍定義

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 01:18:18 +08:00
OG T
e672635edf fix(test): 更新 TestHistoryMessageFormat 適配 Phase 27 雙層策略
Some checks failed
CD Pipeline / build-and-deploy (push) Has been cancelled
Details 
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-10 01:12:00 +08:00