I have noticed during large 3D Flex Train and Reconstruction jobs (hundreds of thousands of particles) that they will report ‘failed’ with the ‘Job is unresponsive - no heartbeat received in 60 seconds’ error, but if you check the log they are actually still running just fine; ie. iterations are continuing to increase. Then, once the job finishes it will properly report ‘job complete’. I am running Cryosparc 4.1.2.
I think this is just a superficial issue, but it would be nice to have 3D Flex jobs report their state properly.