See https://github.com/rr-debugger/rr/actions/runs/9656364835/job/26633773273

I checked out an AWS instance where I could replay the failing trace, but could not reproduce the bug. I don't know exactly what happened, because none of our log files show anything, except debug.err has

2024-06-25T04:20:50.1165441Z 923: process 52475 sent SIGURG

Pid 52475 is rr, as expected, so rr sent a SIGURG presumably due to a fatal abort. But by the time test-monitor gets around to attaching to rr for a stack trace, the rr process has disappeared!