results over the last 7 days — who's slacking?
% of successful probes where the model did NOT return the expected response · lower is better · daily over 7 days
daily averages in your local timezone · only successful probes