I suspect the fan might be broken, the gpu then overheats and shuts down. (Unknown PCI header type ‘127’ (ff)) points to it.
Please run nvidia-bug-report.sh as root and attach the resulting nvidia-bug-report.log.gz file to your post.
This is rather embaressing but… the GPU cable underneath the fan was preventing it from spinning. The workstation is made so that there is no way to visually notice this unless you stick your finger under the card.
Problem solved and stress (fear) levels back to baseline.