feat(p3.2+adr-100): Model Version Tracker + SLO 自治 + KB rot cleaner
Some checks failed
run-migration / migrate (push) Failing after 12s
CD Pipeline / build-and-deploy (push) Has been cancelled

Wave 8 P3.2 模型版本追蹤 + ADR-100 SLO 自我治理 + 配套:

P3.2 — Model Version Tracking:
- model_version_probe.py (268 行) — 探測 Ollama / OpenRouter 等 provider 的 model version
- model_version_tracker.py (101 行) — 對齊 PG provider_version_history 表
- migrations/p3_2_provider_version_history.sql + rollback — 25 行 schema
- db/models.py +32 行 — ProviderVersionHistory ORM

ADR-100 — AI 自主化 SLO:
- docs/adr/ADR-100-ai-autonomous-slo.md (167 行) — 飛輪 SLO 設計與閾值
- ops/monitoring/slo-rules.yml (254 行) — Prometheus SLO recording rules + alerts
- ops/monitoring/tests/test_slo_rules.yaml (242 行) — promtool unit tests

整合修改:
- main.py +72 行 — Lifespan 啟動 model_version_probe + KB rot cleaner schedule
- gitea_webhook.py +45 行 — webhook 接收 model 版本變化通知
- ci_auto_repair.py / evidence_snapshot.py / pre_decision_investigator.py — 配合接線

新測試:
- test_kb_rot_cleaner_schedule.py (120 行) — 9 tests pass
- test_slo_rules.yaml — promtool 驗收

Tests: 9 passed (test_kb_rot_cleaner_schedule)

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Co-Authored-By: Multiple Engineers (P3.2 + ADR-100) <noreply@anthropic.com>
This commit is contained in:
Your Name
2026-04-27 14:54:19 +08:00
parent 9908fdf50d
commit 025a493f06
14 changed files with 1370 additions and 36 deletions

View File

@@ -509,6 +509,51 @@ async def lifespan(_app: FastAPI) -> AsyncGenerator[None, None]:
except Exception as e:
logger.warning("knowledge_decay_loop_schedule_failed", error=str(e))
# ADR-087 Phase 6: KB 腐爛清理(月度)— 每月 1 號 03:00 台北時間
# 掃描 knowledge_entries 中腐爛條目(廢棄 K8s API / Prometheus pattern / 180d 未引用)
# 2026-04-27 P3.1-T3 by Claude
try:
from src.utils.timezone import now_taipei
from datetime import datetime as _dt
async def _run_kb_rot_cleaner_loop() -> None:
from src.jobs.kb_rot_cleaner import get_kb_rot_cleaner
import asyncio as _asyncio
while True:
try:
now = now_taipei()
# 計算下次月初 3 點(台北時間)
if now.day == 1 and now.hour < 3:
next_run = now.replace(hour=3, minute=0, second=0, microsecond=0)
elif now.month == 12:
next_run = now.replace(
year=now.year + 1, month=1, day=1,
hour=3, minute=0, second=0, microsecond=0,
)
else:
next_run = now.replace(
month=now.month + 1, day=1,
hour=3, minute=0, second=0, microsecond=0,
)
sleep_sec = (next_run - now).total_seconds()
logger.info("kb_rot_cleaner_next_run", next_run=next_run.isoformat(), sleep_sec=int(sleep_sec))
await _asyncio.sleep(sleep_sec)
try:
result = await get_kb_rot_cleaner().run()
logger.info("kb_rot_cleaner_completed", stale_count=result.stale_count, total=result.total_scanned)
except Exception as _e:
logger.exception("kb_rot_cleaner_failed", error=str(_e))
except _asyncio.CancelledError:
break
except Exception as _e:
logger.exception("kb_rot_cleaner_loop_error", error=str(_e))
await _asyncio.sleep(3600) # 1h 後重試
asyncio.create_task(_run_kb_rot_cleaner_loop())
logger.info("kb_rot_cleaner_loop_scheduled", trigger="monthly_day1_03h_taipei")
except Exception as e:
logger.warning("kb_rot_cleaner_loop_schedule_failed", error=str(e))
# ADR-083 Phase 3: Fine-tune JSONL 匯出(每週)— EvidenceSnapshot × AgentSession → JSONL
# 2026-04-15 ogt + Claude Sonnet 4.6(亞太): Phase 3 初始建立
try:
@@ -590,6 +635,33 @@ async def lifespan(_app: FastAPI) -> AsyncGenerator[None, None]:
except Exception as e:
logger.warning("ollama_failover_system_start_failed", error=str(e))
# 2026-04-27 P3.2.2 by Claude — AI Provider 版本追蹤(每 1 小時)
# 探測 5 Providerollama/ollama_188/gemini/claude/openclaw_nemo版本
# 寫入 ai_provider_version_history版本變更時 log warningP3.2.3 alerter 後續整合
try:
async def _run_model_version_tracker_loop() -> None:
from src.services.model_version_tracker import get_model_version_tracker
tracker = get_model_version_tracker()
while True:
try:
await asyncio.sleep(3600) # 每 1 小時
result = await tracker.run_probe_cycle()
logger.info(
"model_version_probe_cycle_done",
probed=result["probed"],
changed=result["changed"],
)
except asyncio.CancelledError:
break
except Exception as _loop_err:
logger.exception("model_version_tracker_loop_error", error=str(_loop_err))
await asyncio.sleep(60) # 錯誤後 1 分鐘重試
asyncio.create_task(_run_model_version_tracker_loop())
logger.info("model_version_tracker_scheduled", interval_sec=3600)
except Exception as e:
logger.warning("model_version_tracker_schedule_failed", error=str(e))
yield
# Shutdown