I think you're conflating the scraping of time-series data with it's storage.