mirror of
https://github.com/hansjone/netx.git
synced 2026-10-09 05:30:46 +08:00
oliver
214bc56774
Split the ops wall to its own route with fullscreen charts, persist collect errors as device events, and polish wall legend layout. Co-authored-by: Cursor <cursoragent@cursor.com>
31 lines
966 B
Python
31 lines
966 B
Python
"""Startup recovery for interrupted port traffic collect rounds."""
|
|
|
|
from __future__ import annotations
|
|
|
|
import logging
|
|
from datetime import datetime
|
|
|
|
from sqlalchemy.orm import Session
|
|
|
|
from .models import PortTrafficDevice
|
|
|
|
_log = logging.getLogger("netx.port_traffic.recovery")
|
|
|
|
|
|
def recover_port_traffic_on_startup(db: Session) -> int:
|
|
"""Clear stuck collect_running flags so scheduler can resume."""
|
|
now = datetime.utcnow()
|
|
rows = db.query(PortTrafficDevice).filter(PortTrafficDevice.collect_running.is_(True)).all()
|
|
n = 0
|
|
for device in rows:
|
|
device.collect_running = False
|
|
if not device.last_collect_ended_at:
|
|
device.last_collect_ended_at = now
|
|
if not device.last_error:
|
|
device.last_error = "requeued_after_restart"
|
|
device.updated_at = now
|
|
n += 1
|
|
if n:
|
|
db.commit()
|
|
_log.info("port_traffic recovery cleared collect_running on %s device(s)", n)
|
|
return n
|