(feature request) Drift detection in Kubernetes operator versioned workflow #3664
Description
Not really an issue, but more like a feature request. It would be very useful for the Atlas k8s operator to perform some pre-migration drift detection check before applying new migrations that appeared in the ConfigMap.
What I mean exactly is the operator could run atlas schema diff on the migration directory for all committed migrations and the target database before applying newly appeared migration files. This FAQ is a good example of the flow I'm taking about but ideally, especially in k8s-based environments, the CI shouldn't have access to the database which makes this proposition/workaround impossible.
Perhaps we should add a flag like --strict or --drift-check to atlas migrate apply that would first check for drift before applying the new migrations OR some option for that in the k8s operator itself? That way the migration would fail if drift is detected and in the k8s world we could do some alerting and stuff.