[rust/rqd] start_time timestamp incorrect format (second instead of microsecond)

[aoblet]

Describe the bug
When using timeout feature on Layer, dispatcher is always killing the task.
This is due to incorrect RunningFrameInfo:start_time format being sent by rqd via RunningFrame.clone_into_running_frame_info
Instead of being in microsecond format, second format is sent.
Then the condition check on dispatcher side is never correct.

To Reproduce
Steps to reproduce the behavior:

1. Submit a job with a timeout on Layer
2. Launch a worker with rust rqd
3. See exception raised "Frame timed out" in the log (worker and dispatcher)

Expected behavior
The task not to be killed if the timeout is respected.

Version Number
1.13.8 - 95d93c2