This article explains that a FPC could crash because of a race condition. Because of this, FPC will be down and traffic will be affected. This is an unusual situation.

In chassisd we observed that FPC rebooted due to NMI timeout.

In messages it was observed "DAEMON-3-CHASSISD_IPC_CONNECTION_DROPPED", which could indicate that the communication between the Routing Engine (RE) and the FPC was interrupted. This could be caused by a variety of factors, including FPC resets, resource exhaustion, software bugs, or hardware failures.

We reviewed the VMHOST logs and discovered that a kernel exception occurred at the time of the issue. Also a FPC Core was observed. 

On all Junos and Junos Evolved platforms, in a rare scenario, the FPC will go down due to core.

 PR1739595 describes this issue.