ayrisdev/crawl4ai
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add cdp_cleanup_on_close flag to prevent memory leaks in cloud/server scenarios

Browse Source
This commit is contained in:
unclecode
2025-12-13 06:25:25 +00:00
parent 8ae908bede
commit 66941a59e8
2 changed files with 13 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
crawl4ai/async_configs.py
View File

@@ -380,6 +380,10 @@ class BrowserConfig:
target_id (str or None): Pre-existing CDP target ID (page) to use. When provided along with target_id (str or None): Pre-existing CDP target ID (page) to use. When provided along with
browser_context_id, the crawler will reuse this target instead of creating browser_context_id, the crawler will reuse this target instead of creating
a new page. Default: None. a new page. Default: None.
cdp_cleanup_on_close (bool): When True and using cdp_url, the close() method will still clean up
the local Playwright client resources. Useful for cloud/server scenarios
where you don't own the remote browser but need to prevent memory leaks
from accumulated Playwright instances. Default: False.
debugging_port (int): Port for the browser debugging protocol. Default: 9222. debugging_port (int): Port for the browser debugging protocol. Default: 9222.
use_persistent_context (bool): Use a persistent browser context (like a persistent profile). use_persistent_context (bool): Use a persistent browser context (like a persistent profile).
Automatically sets use_managed_browser=True. Default: False. Automatically sets use_managed_browser=True. Default: False.
@@ -436,6 +440,7 @@ class BrowserConfig:
cdp_url: str = None, cdp_url: str = None,
browser_context_id: str = None, browser_context_id: str = None,
target_id: str = None, target_id: str = None,
cdp_cleanup_on_close: bool = False,
use_persistent_context: bool = False, use_persistent_context: bool = False,
user_data_dir: str = None, user_data_dir: str = None,
chrome_channel: str = "chromium", chrome_channel: str = "chromium",
@@ -477,6 +482,7 @@ class BrowserConfig:
self.cdp_url = cdp_url self.cdp_url = cdp_url
self.browser_context_id = browser_context_id self.browser_context_id = browser_context_id
self.target_id = target_id self.target_id = target_id
self.cdp_cleanup_on_close = cdp_cleanup_on_close
self.use_persistent_context = use_persistent_context self.use_persistent_context = use_persistent_context
self.user_data_dir = user_data_dir self.user_data_dir = user_data_dir
self.chrome_channel = chrome_channel or self.browser_type or "chromium" self.chrome_channel = chrome_channel or self.browser_type or "chromium"

6
crawl4ai/browser_manager.py
View File

@@ -1234,6 +1234,12 @@ class BrowserManager:
async def close(self): async def close(self):
"""Close all browser resources and clean up.""" """Close all browser resources and clean up."""
if self.config.cdp_url: if self.config.cdp_url:
# When using external CDP, we don't own the browser so skip most cleanup.
# But if cdp_cleanup_on_close is True, still clean up local Playwright client
# to prevent memory leaks in server/cloud scenarios.
if self.config.cdp_cleanup_on_close and self.playwright:
await self.playwright.stop()
self.playwright = None
return return
if self.config.sleep_on_close: if self.config.sleep_on_close: