Good post, breaks down the threat models really well. =) One small note, I think there may be a small issue in one of the code blocks explaining the client key

Good post, breaks down the threat models really well. =) One small note, I think there may be a small issue in one of the code blocks explaining the client key: > cache_namespace = tenant_id || client_secret I believe that should say `tenant_id + client_secret`, ya? To append them together? If you OR them out it'll default to the `tenant_id`, which isn't what you want in this case, I believe.