This can be caused by non-reliable selector of the targeted table.
Sometimes it is working, sometimes it does not. If you are scraping a public website, this can be even a built-in protection against scraping.
Search for “Selector” in properties of Extract Structured Data. Open that selector in UiExplorer, try to refresh the page and see if it is still pointing at the correct target.
you can play around with this property in case we have the feeling that the table gets scraped sooner than it loads. this may help:
You can share your workflow and/or the selector of that activity with us, that will help.