Thermal threshold. Modern server equipment is designed to shut itself off to prevent damage to the hardware under a number of conditions. In this case, with a faulting fan, the most likely cause is that the server's internals are getting too hot, and it is shutting down to prevent melting circuit boards and pesky electrical fires... either of which would cause the server to power off in much less controlled, more spectacular and more expensive fashion anyway.
(The other common cause would be electrical component failures - severe voltage variances, blown capacitors, etc., but that seems like the less likely cause here.)
This information is generally logged by the system, and assuming you have an iDRAC subsystem in this server, you'd login to it to verify the cause is a thermal threshold being surpassed, which is just a technical way of saying something in your server is getting too hot to operate safely.
The obvious remedy would be to replace the faulty fan, but there may be other solutions as well, like unplugging unneeded components, improving airflow to the server, providing cooled intake air or even blowing out dust with canned air (dust is a surprisingly good thermal insulator and will cause your components to retain heat).