fix(browser): replace expired cloud sessions

This commit is contained in:
Gille 2026-08-01 13:11:28 -06:00 committed by kshitij
parent 927662e48f
commit 58e85f4314
4 changed files with 169 additions and 7 deletions

View File

@ -26,6 +26,7 @@ Session metadata contract (preserved from the legacy ``CloudBrowserProvider``)::
"session_name": str, # unique name for agent-browser --session
"bb_session_id": str, # provider session ID (for close/cleanup)
"cdp_url": str, # CDP websocket URL
"expires_at": str, # optional provider-authoritative ISO timestamp
"features": dict, # feature flags that were enabled
"external_call_id": str, # optional, managed-gateway billing key
}
@ -96,6 +97,7 @@ class BrowserProvider(abc.ABC):
"session_name": str, # unique name for agent-browser --session
"bb_session_id": str, # provider session ID (for close/cleanup)
"cdp_url": str, # CDP websocket URL
"expires_at": str, # optional provider-authoritative ISO timestamp
"features": dict, # feature flags that were enabled
}

View File

@ -246,6 +246,10 @@ class BrowserUseBrowserProvider(BrowserProvider):
"session_name": session_name,
"bb_session_id": session_data["id"],
"cdp_url": cdp_url,
# Browser Use sessions have a fixed server-side lifetime. Preserve
# the authority returned by the API so the dispatcher can retire an
# expired CDP endpoint instead of reconnecting to it indefinitely.
"expires_at": session_data.get("timeoutAt"),
"features": {"browser_use": True},
"external_call_id": external_call_id,
}

View File

@ -0,0 +1,93 @@
"""Regression coverage for provider-authoritative cloud browser expiry."""
from unittest.mock import Mock
import tools.browser_tool as browser_tool
from plugins.browser.browser_use import provider as browser_use_provider
def _isolate_browser_state(monkeypatch):
monkeypatch.setattr(browser_tool, "_active_sessions", {})
monkeypatch.setattr(browser_tool, "_session_last_activity", {})
monkeypatch.setattr(browser_tool, "_start_browser_cleanup_thread", lambda: None)
monkeypatch.setattr(browser_tool, "_ensure_cdp_supervisor", lambda task_id: None)
def test_browser_use_preserves_provider_timeout(monkeypatch):
provider = browser_use_provider.BrowserUseBrowserProvider()
response = Mock(
ok=True,
headers={},
)
response.json.return_value = {
"id": "browser-session-1",
"cdpUrl": "ws://browser-use.example/devtools/browser/1",
"timeoutAt": "2030-01-01T00:05:00Z",
}
monkeypatch.setattr(
provider,
"_get_config",
lambda: {
"api_key": "test-key",
"base_url": "https://api.browser-use.example/api/v3",
"managed_mode": False,
},
)
monkeypatch.setattr(browser_use_provider.requests, "post", Mock(return_value=response))
session = provider.create_session("task-1")
assert session["expires_at"] == "2030-01-01T00:05:00Z"
def test_live_cloud_session_is_reused(monkeypatch):
_isolate_browser_state(monkeypatch)
existing = {
"session_name": "existing",
"bb_session_id": "browser-session-1",
"cdp_url": "ws://browser-use.example/devtools/browser/1",
"expires_at": "2999-01-01T00:05:00Z",
}
browser_tool._active_sessions["task-1"] = existing
provider = Mock()
monkeypatch.setattr(browser_tool, "_get_cloud_provider", lambda: provider)
session = browser_tool._get_session_info("task-1")
assert session is existing
provider.create_session.assert_not_called()
def test_expired_cloud_session_is_replaced_without_reusing_dead_cdp(monkeypatch):
_isolate_browser_state(monkeypatch)
browser_tool._active_sessions["task-1"] = {
"session_name": "expired",
"bb_session_id": "browser-session-old",
"cdp_url": "ws://browser-use.example/devtools/browser/old",
"expires_at": "2020-01-01T00:05:00Z",
}
browser_tool._session_last_activity["task-1"] = 1.0
provider = Mock()
provider.create_session.return_value = {
"session_name": "replacement",
"bb_session_id": "browser-session-new",
"cdp_url": "ws://browser-use.example/devtools/browser/new",
"expires_at": "2999-01-01T00:05:00Z",
}
monkeypatch.setattr(browser_tool, "_get_cloud_provider", lambda: provider)
monkeypatch.setattr(browser_tool, "_get_cdp_override", lambda: "")
monkeypatch.setattr(browser_tool, "_stop_cdp_supervisor", Mock())
monkeypatch.setattr(browser_tool, "_maybe_stop_recording", Mock())
monkeypatch.setattr(browser_tool, "_run_browser_command", Mock())
monkeypatch.setattr(browser_tool.os.path, "exists", lambda path: False)
session = browser_tool._get_session_info("task-1")
assert session["bb_session_id"] == "browser-session-new"
assert browser_tool._active_sessions["task-1"] is session
assert "task-1" in browser_tool._session_last_activity
provider.close_session.assert_called_once_with("browser-session-old")
provider.create_session.assert_called_once_with("task-1")
browser_tool._run_browser_command.assert_not_called()

View File

@ -61,6 +61,7 @@ import sys
import tempfile
import threading
import time
from datetime import datetime, timezone
from typing import Dict, Any, Optional, List, Tuple, Union
from pathlib import Path
from agent.redact import redact_cdp_url
@ -1568,6 +1569,42 @@ _cleanup_running = False
_cleanup_lock = threading.Lock()
def _session_expiry_timestamp(session_info: Dict[str, Any]) -> Optional[float]:
"""Return a provider-authoritative session expiry as epoch seconds.
Cloud providers may omit ``expires_at``. Unknown or malformed values are
therefore treated as having no known expiry, preserving the existing
lifecycle for local browsers and providers without an expiry contract.
"""
value = session_info.get("expires_at")
if isinstance(value, (int, float)) and not isinstance(value, bool):
return float(value)
if not isinstance(value, str) or not value.strip():
return None
normalized = value.strip()
if normalized.endswith(("Z", "z")):
normalized = f"{normalized[:-1]}+00:00"
try:
parsed = datetime.fromisoformat(normalized)
except ValueError:
logger.warning("Ignoring invalid cloud browser session expiry timestamp")
return None
if parsed.tzinfo is None:
parsed = parsed.replace(tzinfo=timezone.utc)
return parsed.timestamp()
def _session_has_expired(
session_info: Dict[str, Any], *, now: Optional[float] = None
) -> bool:
"""Return whether a cached browser session crossed its provider deadline."""
expires_at = _session_expiry_timestamp(session_info)
if expires_at is None:
return False
return (time.time() if now is None else now) >= expires_at
def _emergency_cleanup_all_sessions():
"""
Emergency cleanup of all active browser sessions.
@ -2152,7 +2189,22 @@ def _get_session_info(task_id: Optional[str] = None) -> Dict[str, Any]:
with _cleanup_lock:
# Check if we already have a session for this task
if task_id in _active_sessions:
return _active_sessions[task_id]
existing_session = _active_sessions[task_id]
else:
existing_session = None
if existing_session is not None:
if not _session_has_expired(existing_session):
return existing_session
logger.info(
"Replacing expired cloud browser session for task %s",
task_id,
)
_cleanup_single_browser_session(task_id)
# Cleanup removes the activity entry. The replacement session must be
# tracked by the inactivity reaper just like an initial session.
_update_session_activity(task_id)
# Hybrid routing: session keys ending with ``::local`` force a local
# Chromium regardless of the globally-configured cloud provider. Public
@ -4539,12 +4591,23 @@ def _cleanup_single_browser_session(task_id: str) -> None:
# Stop auto-recording before closing (saves the file)
_maybe_stop_recording(task_id)
# Try to close via agent-browser first (needs session in _active_sessions)
try:
_run_browser_command(task_id, "close", [], timeout=10)
logger.debug("agent-browser close command completed for task %s", task_id)
except Exception as e:
logger.warning("agent-browser close failed for task %s: %s", task_id, e)
# An expired cloud CDP URL cannot accept an agent-browser close command.
# Avoid feeding it back through _get_session_info(), which would try to
# renew the session recursively while cleanup is still in progress.
if _session_has_expired(session_info):
logger.debug(
"Skipping agent-browser close for expired session %s",
task_id,
)
else:
try:
_run_browser_command(task_id, "close", [], timeout=10)
logger.debug(
"agent-browser close command completed for task %s",
task_id,
)
except Exception as e:
logger.warning("agent-browser close failed for task %s: %s", task_id, e)
# Now remove from tracking under lock
with _cleanup_lock: