fix: handle unique constraint violation in manual fetch endpoint

When a duplicate timestamp occurs (rare but possible), return the
existing record instead of failing with a 500 error. This matches
the worker's ON CONFLICT DO NOTHING behavior.

Added test for duplicate timestamp handling.

backend/routes/audit.py

@@ -6,6 +6,7 @@ from typing import TypeVar
 from fastapi import APIRouter, Depends, Query
 from pydantic import BaseModel
 from sqlalchemy import desc, func, select
+from sqlalchemy.exc import IntegrityError
 from sqlalchemy.ext.asyncio import AsyncSession
 
 from auth import require_permission
@@ -200,8 +201,20 @@ async def fetch_price_now(
         timestamp=timestamp,
     )
     db.add(record)
-    await db.commit()
-    await db.refresh(record)
+
+    try:
+        await db.commit()
+        await db.refresh(record)
+    except IntegrityError:
+        # Duplicate timestamp - return the existing record
+        await db.rollback()
+        query = select(PriceHistory).where(
+            PriceHistory.source == SOURCE_BITFINEX,
+            PriceHistory.pair == PAIR_BTC_EUR,
+            PriceHistory.timestamp == timestamp,
+        )
+        result = await db.execute(query)
+        record = result.scalar_one()
 
     return PriceHistoryResponse(
         id=record.id,