in cases where cronjob controller goes offline for a prolongued period of time new jobs cannot start since kubernetes assumes that there were over a 100 failures (which is a default threshold) and refuses to start new jobs at all. Setting
startingDeadlineSeconds to a reasonable length in seconds allows to overcome that limitation. So, with cron set to run every 5 min we shouln't see more 100 failures within, say 200min (200*60 == 12000). Setting it to a even lower number basically ensures that jobs will always run even if cluster/controler was offline for a longer period of time.