feat(09-01): implement retention cleanup service with configurable retention period

- Add CONFIG_RETENTION_DAYS setting (default 90) to config.py
- Create retention_service.py with cleanup_expired_snapshots (parameterized SQL via make_interval)
- APScheduler IntervalTrigger runs cleanup every 24h with 1h jitter
- Prometheus counter and histogram for observability
- CASCADE FKs handle diff/change deletion automatically
- All 4 unit tests pass

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

backend/app/config.py

@@ -125,6 +125,9 @@ class Settings(BaseSettings):
     PASSWORD_RESET_TOKEN_EXPIRE_MINUTES: int = 30
     APP_BASE_URL: str = "http://localhost:3000"
 
+    # Retention cleanup — delete config snapshots older than N days
+    CONFIG_RETENTION_DAYS: int = 90
+
     # App settings
     APP_NAME: str = "TOD - The Other Dude"
     APP_VERSION: str = "0.1.0"

backend/app/services/retention_service.py

@@ -0,0 +1,85 @@
+"""Retention cleanup service — deletes config snapshots older than CONFIG_RETENTION_DAYS.
+
+Runs as an APScheduler IntervalTrigger job (every 24h). CASCADE FK constraints
+on router_config_diffs and router_config_changes handle associated data automatically.
+"""
+
+import logging
+from typing import Optional
+
+from apscheduler.schedulers.asyncio import AsyncIOScheduler
+from apscheduler.triggers.interval import IntervalTrigger
+from prometheus_client import Counter, Histogram
+from sqlalchemy import text
+
+from app.config import settings
+from app.database import AdminAsyncSessionLocal
+
+logger = logging.getLogger(__name__)
+
+_scheduler: Optional[AsyncIOScheduler] = None
+
+# Prometheus metrics
+config_snapshots_cleaned_total = Counter(
+    "config_snapshots_cleaned_total",
+    "Cumulative count of expired config snapshots deleted by retention cleanup",
+)
+config_retention_cleanup_duration_seconds = Histogram(
+    "config_retention_cleanup_duration_seconds",
+    "Duration of retention cleanup execution",
+)
+
+
+async def cleanup_expired_snapshots() -> int:
+    """Delete config snapshots older than CONFIG_RETENTION_DAYS.
+
+    CASCADE FK constraints on router_config_diffs and router_config_changes
+    automatically remove associated rows.
+
+    Returns the number of deleted snapshots.
+    """
+    days = settings.CONFIG_RETENTION_DAYS
+
+    with config_retention_cleanup_duration_seconds.time():
+        async with AdminAsyncSessionLocal() as session:
+            result = await session.execute(
+                text(
+                    "DELETE FROM router_config_snapshots "
+                    "WHERE collected_at < NOW() - make_interval(days => :days)"
+                ),
+                {"days": days},
+            )
+            await session.commit()
+            deleted = result.rowcount
+
+    config_snapshots_cleaned_total.inc(deleted)
+    logger.info("retention cleanup complete", extra={"deleted_snapshots": deleted, "retention_days": days})
+    return deleted
+
+
+async def start_retention_scheduler() -> None:
+    """Start APScheduler with a 24-hour interval job for retention cleanup."""
+    global _scheduler
+    _scheduler = AsyncIOScheduler(timezone="UTC")
+    _scheduler.add_job(
+        cleanup_expired_snapshots,
+        trigger=IntervalTrigger(hours=24, jitter=3600),
+        id="retention_cleanup",
+        name="Config snapshot retention cleanup",
+        max_instances=1,
+        replace_existing=True,
+    )
+    _scheduler.start()
+    logger.info(
+        "retention scheduler started (every 24h, retention_days=%d)",
+        settings.CONFIG_RETENTION_DAYS,
+    )
+
+
+async def stop_retention_scheduler() -> None:
+    """Gracefully shutdown the retention scheduler."""
+    global _scheduler
+    if _scheduler:
+        _scheduler.shutdown(wait=False)
+        _scheduler = None
+        logger.info("retention scheduler stopped")