Pyrocumulonimbus (pyroCb) cloud formation is a form of extreme fire behavior which can alter weather conditions around the fire and lead to long-range smoke transport and climate impacts. PyroCbs also create extremely challenging conditions for observing the underlying fire activity, both from the perspective of airborne mission safety and optically thick plumes obstructing satellite-based fire detection. The lack of active fire data during pyroCbs leads to underestimation of fire radiative power (FRP)-based smoke emissions estimates, and makes it particularly challenging to track fire spread and behavior. In this study, we investigate VIIRS active fire information during pyroCb events, leveraging a catalog of 145 pyroCbs from summer 2023 in Canada. This set of events is particularly data-rich, given the availability of a third VIIRS sensor in 2023 as well as more frequent coverage from overlapping VIIRS swaths at high latitudes. Focusing our analysis on 45 fires that produced pyroCbs observed by VIIRS (i.e., with maximum pyroCb altitude within 1 hour of one or more VIIRS overpasses), we found a decrease of approximately 15-20% in active fire detections during the afternoon of the pyroCb relative to the adjacent nighttime overpasses (+/- 12 hours). This decrease represents a significant gap in data availability due to pyroCb plumes when fire activity and FRP should otherwise be at a maximum. Next, we increased the quantity of available fire information by including lower-intensity thermal anomalies labeled as “candidate” fire pixels in the VIIRS fire algorithm. These pixels represent hotspots that were previously excluded due to slightly lower intensity and/or cloud masking. Including candidate fires boosted the total number of fire detections by 53% across the +/- 4 days surrounding each pyroCb event, and by 84-103% (roughly doubling) in the hour surrounding the maximum pyroCb time. Finally, we apply our dataset, leveraging the high frequency of VIIRS observations plus the additional candidate detections, to improve fire perimeter tracking and FRP estimates over time, and to relate this extreme fire behavior to potential drivers such as fuels and weather.