most teams just run a terraform plan on a cron job and scream when the output isnt empty. this is
not actual drift detection, its just
noise from every tiny change. if your alerts are basically
useless spam at this point , you might need to rethink how you handle state. does anyone actually use a real reconciliation loop instead of just pinging slack?
more here:
https://dzone.com/articles/drift-detection-with-severity-classification