Measuring
What you are actually taking
The module's screen shows the last seven days of failures, broken down by reason: bot trap triggered, submitted too fast, wrong answer, challenge expired, token replayed, token forged. That breakdown tells you what you are dealing with — a spike in token replayed is not handled like a spike in wrong answer.
Two maintenance actions sit on the same screen: purging expired token fingerprints, and renewing the signing key — the latter invalidating every challenge in flight straight away.