For example, I would think at quickest a cron job would run every minute which means if a post was made and deleted after a scrape already took place but before the next scrape was due (within 59 seconds), it will not be archived.
You wouldn't start a new cronjob for every scrape, and even if you did - there are ways to start them more often than once a minute (e.g. multiple jobs with different sleep intervals at the beginning). Anyway, the point is that scrapers can catch new posts within seconds.