How would Trains go about sending any missing data to the server in the following scenarios?
Disclaimer: I'm part of the allegro.ai Trains team
sdk.network.iteration.retry_backoff_factor_sec
parameter in your ~/trains.conf
file, see example here