Docker Container shows different CPU Usage with different tools
It's not exactly clear in the node exporter what instance or container you are monitoring, though it seems like the node exporter it showing the total machine CPU usage in 0-100 format and ctop shows in 100% per vCPU format.Also try using docker stats, that should show all running containers resource usage, from cpu to network and disk usage, when using it each vCPU would as 100% so your total will be 1600% for 16 vCPU.
Regarding the cAdvisor output it doesn't show the same time range as the grafana node exporter so it would be hard to make a hard conclusion but it seems like similarly to ctop and docker stats it shows on a per core basis but instead of percentage it shows in 'cores' unit of measurement