liveBook · Manning

OnFailAction

As explained in Machine Learning Platform Engineering

OnFailAction is a Guardrails AI failure-handling setting that specifies what a validator does when content violates its criteria.

OnFailAction controls how a Guardrails AI validator responds when content violates its criteria. Choose the response according to risk and whether the content can be safely corrected. section 13.3.1

Available strategies include:

  • EXCEPTION or REFRAIN blocks processing for high-risk content. section 13.3.1
  • FIX or FILTER sanitizes content and continues validation. section 13.3.1
  • REASK or FIX_REASK requests regenerated content when improvement is possible. section 13.3.1
  • NOOP logs the failure without intervening. section 13.3.1

Risk-based input validation

Configure CompetitorCheck with OnFailAction.EXCEPTION to stop competitor discussions, and GuardrailsPII with OnFailAction.FIX to redact email addresses, credit cards, or SSNs before processing continues.

How Machine Learning Platform Engineering covers it

How deep it goes
Explains seven response strategies and contrasts blocking, sanitizing, regenerating, and logging, without detailing their internals.
Examples
Uses competitor-name blocking, PII redaction, and a Guardrails AI Python configuration.
What it assumes
Assumes familiarity with Guardrails AI validators and input validation while teaching failure choices through configuration.
What you can do after
Apply risk-based failure strategies when configuring Guardrails AI validators.
How it connects
Connects validators with input validation, PII detection, content filtering, and safety guardrails.
Cover of Machine Learning Platform Engineering
liveBookmanning.com
Read Machine Learning Platform Engineering on liveBook Get the book at manning.com

This explainer was generated by Manning's knowledge graph from the text of Machine Learning Platform Engineering. The section links point to the parts of the book it draws on.