Fifty parallel scrapers, one shared logged-in session

Vendor portal. Two-factor login. The login takes 90 seconds and you'd rather not do it fifty times in parallel: they'd notice. But you have 50,000 records to fetch and one worker would take three days.

You log in once with session_id: "vendor-portal". The session: cookies the portal set: sits in hoody-curl's session store. You spin up fifty container-side scraper workers, each firing curl requests at vendor pages with session_id: "vendor-portal". Every request carries the cookies the login established.

When the cookies expire, you re-login once. You don't restart fifty workers. The session is decoupled from the caller.

Login once. Scrape from anywhere. Cookies follow the session id, not the worker.

Cookies follow the session id, not the worker.