Making scalable source control is much easier when you can rely on a distributed database (Spanner) and object storage (GCS/S3/...)!