From 7e1a4699e9bd71390ab823cb41fafeb9af67232e Mon Sep 17 00:00:00 2001
From: I Luk Kim <yirugi@gmail.com>
Date: Tue, 14 Apr 2026 22:00:15 -0700
Subject: [PATCH] =?UTF-8?q?fix:=20request=5Flogs=2018GB=20=EB=B8=94?=
 =?UTF-8?q?=EB=A1=9C=ED=8A=B8=20=EB=B0=A9=EC=A7=80=20=E2=80=94=207?=
 =?UTF-8?q?=EC=9D=BC=20retention=20+=20autovacuum=20=EA=B0=95=ED=99=94?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

원인: request_logs에 cleanup 정책이 없어 65M행이 쌓였다가 DELETE되면서
VACUUM FULL 없이 18GB dead page bloat 발생.
PostgreSQL 메모리/디스크 압박 → API worker OOM kill의 간접 원인.

수정:
- _request_log_cleanup() 백그라운드 코루틴 추가
  → 1시간마다 실행, 7일 초과 request_logs 자동 DELETE
  → start_request_log_flusher()에서 함께 스케줄
- request_logs autovacuum scale_factor 0.2 → 0.01
  → 1% 변경 시 autovacuum 즉시 실행, dead tuple 빠르게 회수

즉각 조치: TRUNCATE request_logs 실행 (25GB → 56kB)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
---
 app/middleware/error_logger.py | 26 +++++++++++++++++++++++++-
 1 file changed, 25 insertions(+), 1 deletion(-)

diff --git a/app/middleware/error_logger.py b/app/middleware/error_logger.py
index 93a8f46..a8198a9 100644
--- a/app/middleware/error_logger.py
+++ b/app/middleware/error_logger.py
@@ -16,7 +16,7 @@ import json
 import time
 import traceback
 import uuid
-from datetime import datetime, timezone
+from datetime import datetime, timedelta, timezone
 from typing import Callable, Optional
 
 from fastapi import Request, Response
@@ -38,6 +38,8 @@ logger = logging.getLogger(__name__)
 _REQUEST_LOG_QUEUE: asyncio.Queue = asyncio.Queue(maxsize=10_000)
 _FLUSH_INTERVAL_SECONDS: float = 2.0   # flush at most every 2 s
 _FLUSH_BATCH_SIZE: int = 500           # or when 500 entries are queued
+_LOG_RETENTION_DAYS: int = 7           # delete request_logs older than this
+_CLEANUP_INTERVAL_HOURS: int = 1       # run cleanup every N hours
 
 
 async def _flush_request_logs(entries: list) -> None:
@@ -82,9 +84,31 @@ async def _request_log_flusher() -> None:
             await _flush_request_logs(entries)
 
 
+async def _request_log_cleanup() -> None:
+    """Background coroutine: delete request_logs older than _LOG_RETENTION_DAYS."""
+    while True:
+        await asyncio.sleep(_CLEANUP_INTERVAL_HOURS * 3600)
+        try:
+            cutoff = datetime.now(timezone.utc) - timedelta(days=_LOG_RETENTION_DAYS)
+            from app.core.database import AsyncSessionLocal
+            from sqlalchemy import text
+            async with AsyncSessionLocal() as db:
+                result = await db.execute(
+                    text("DELETE FROM request_logs WHERE created_at < :cutoff"),
+                    {"cutoff": cutoff},
+                )
+                deleted = result.rowcount
+                await db.commit()
+            if deleted:
+                logger.info(f"request_logs cleanup: deleted {deleted} rows older than {_LOG_RETENTION_DAYS}d")
+        except Exception as e:
+            logger.error(f"request_logs cleanup failed: {e}")
+
+
 def start_request_log_flusher() -> None:
     """Schedule the background flusher coroutine.  Call once at app startup."""
     asyncio.ensure_future(_request_log_flusher())
+    asyncio.ensure_future(_request_log_cleanup())
     logger.info("Request log flusher background task started")