Scraper: force-reset endpoint + watchdog to prevent stuck lock

Root cause: threading.Lock held indefinitely when Playwright browser
hangs inside run_in_executor (finally never fires from the async side).

Fixes:
- _acquire_scrape_lock/_release_scrape_lock track monotonic timestamp
- POST /competitors/scrape/reset force-releases the lock and marks any
  running batch as interrupted (queue rows stay intact for retry)
- GET /competitors/status now includes lock_held_seconds
- APScheduler watchdog job every 30 min auto-releases if held >3h
- Settings → Scraper Proxy tab shows live lock status (green/amber)
  with a Force Reset button requiring confirmation

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
jtricerolph 2026-07-09 16:56:35 +00:00
parent d737940a00
commit e781b1e8b9
4 changed files with 178 additions and 9 deletions

View file

@ -14,6 +14,7 @@ Features:
import asyncio
import logging
import threading
import time
import uuid
from datetime import date, timedelta
from decimal import Decimal
@ -32,6 +33,48 @@ logger = logging.getLogger(__name__)
# other of memory (page timeouts → partial results) and double the request
# rate at Booking.com. Guards both manual scrapes and queue processing.
SCRAPE_LOCK = threading.Lock()
_lock_acquired_at: Optional[float] = None # monotonic timestamp when lock was acquired
def _acquire_scrape_lock() -> bool:
global _lock_acquired_at
if SCRAPE_LOCK.acquire(blocking=False):
_lock_acquired_at = time.monotonic()
return True
return False
def _release_scrape_lock():
global _lock_acquired_at
_lock_acquired_at = None
try:
SCRAPE_LOCK.release()
except RuntimeError:
pass # already released — race between watchdog and normal release
def get_lock_status() -> dict:
"""Return lock state and how long it has been held (seconds)."""
locked = SCRAPE_LOCK.locked()
held = int(time.monotonic() - _lock_acquired_at) if (locked and _lock_acquired_at) else None
return {"locked": locked, "held_seconds": held}
def force_reset_scraper(db: Session) -> dict:
"""Force-release the scrape lock and tidy up any stuck DB state.
Safe to call even when the lock is not held."""
was_locked = SCRAPE_LOCK.locked()
_release_scrape_lock()
# Mark any batch left in 'running' state as interrupted
db.execute(text("""
UPDATE booking_scrape_log
SET status = 'interrupted', completed_at = NOW(),
error_message = 'Force-reset by admin'
WHERE status = 'running'
"""))
db.commit()
logger.warning(f"Scrape lock force-reset (was_locked={was_locked})")
return {"was_locked": was_locked, "reset": True}
def get_scraper_backend(db: Session) -> ScraperBackend:
@ -509,7 +552,7 @@ async def run_manual_scrape(
if to_date is None:
to_date = from_date
if not SCRAPE_LOCK.acquire(blocking=False):
if not _acquire_scrape_lock():
logger.warning(f"Manual scrape {from_date}..{to_date} refused — another scrape is running")
return {
'success': False,
@ -519,7 +562,7 @@ async def run_manual_scrape(
try:
return await _run_manual_scrape_locked(db, from_date, to_date)
finally:
SCRAPE_LOCK.release()
_release_scrape_lock()
async def _run_manual_scrape_locked(
@ -694,7 +737,7 @@ async def process_queue(db: Session) -> Dict[str, Any]:
Returns:
Dict with processing results
"""
if not SCRAPE_LOCK.acquire(blocking=False):
if not _acquire_scrape_lock():
logger.warning("Queue processing skipped — another scrape is running")
return {
'success': False,
@ -704,7 +747,7 @@ async def process_queue(db: Session) -> Dict[str, Any]:
try:
return await _process_queue_locked(db)
finally:
SCRAPE_LOCK.release()
_release_scrape_lock()
async def _process_queue_locked(db: Session) -> Dict[str, Any]: