Gen2 dataflows and deployment pipelines still dont seem to work well together

anyone who has worked with Gen2 dataflows and lakehouses in combination with a deployment pipeline probably knows about all the problem this combination had in the past but even with the current implementation there still seem to be problems. We have a lot of Gen2 dataflows which run in a specific order thanks to a data pipeline that handles all of this and it runs perfectly fine on DEV. All of the data is written to lakehouses and some of the dataflows later in the pipeline also have other lakehouses as their source. All of the ones where it matters are set to use the current workspace as the