Error handling for the steps in the extractors are all handled per step. You choose how to handle errors by editing a step and opening the "General" tab in the step options:
Here you'll find several general options for how the step behaves - scroll down till you see "Error handling":
There are 3 types of error handling available:
Stop and fail
This is the default error mode - whenever a step fails with this error mode the entire robot will stop and report the failure. You'll get "Jump to error" options, a screenshot and access to logs to help you figure out why the robot failed.
You can also set a custom error message for when this happens - to be able to add more automatic handling of the error if it is expected. Note that writing a custom error message will completely replace the original error message and might hide the actual explanation for the error - so it's something we recommend using when you're familiar with the platform.
Ignore and continue
This error mode does exactly as advertised - it simply ignores the error and continues as if the step did not fail. Be careful of using this too much - since errors will help you more easily monitor how your robots are doing. Overusing this can hide that your robots are in fact no longer working as expected - so only apply this to things that you expect might fail and is not required for your robot to continue to produce value.
Skip current branch
This error mode will jump to the next branch or iteration of a loop (which ever is closest to the left of the step) and continue from there. Use this when you are iterating a list of things and know you might want to skip certain ones if they fail - or if you are using branches to handle different conditions on the website.
To learn more about branching be sure to read this article about the basics of the extractor.