fix(ai): remove 188 ollama provider
All checks were successful
Code Review / ai-code-review (push) Successful in 12s
CD Pipeline / tests (push) Successful in 1m13s
CD Pipeline / build-and-deploy (push) Successful in 3m36s
CD Pipeline / post-deploy-checks (push) Successful in 1m20s

This commit is contained in:
Your Name
2026-05-06 14:33:16 +08:00
parent 578bf3bc7c
commit 4111ea4f9f
33 changed files with 193 additions and 233 deletions

View File

@@ -4,7 +4,7 @@ LLM Playbook Generator - ADR-104 T1/T2/T6
從成功修復案例生成可治理的 Playbook 草稿。
設計重點:
- 只用 local provider 順序Ollama 111 -> Ollama 188),避免新增雲端成本。
- 只用 local/provider pool 順序GCP-A -> 111 local),避免新增雲端成本。
- LLM 產出必須經 Pydantic + action_parser 安全收斂。
- 不直接 APPROVED先 DRAFT/REVIEW再交治理 job 晉級。
"""
@@ -30,7 +30,6 @@ from src.models.playbook import (
RiskLevel,
SymptomPattern,
)
from src.services.action_parser import is_safe_kubectl_action
from src.services.action_parser import kubectl_safety_reason
logger = structlog.get_logger(__name__)
@@ -218,7 +217,7 @@ class LLMPlaybookGenerator:
executor = get_ai_executor()
result = await executor.execute(
prompt=prompt,
provider_order=["ollama", "ollama_188"],
provider_order=["ollama", "ollama_local"],
context=context,
cache_ttl=86400,
require_local=True,