You are working on a project that requires creating data pipelines to move and process data. You have been asked to scale resources based on the workload. What considerations should you take into account when scaling resources in a data pipeline?