Welcome @grwtz
Memory error messages from tools are referring to the tool execution itself running out of resources. This is usually due to characteristics of the input dataset (problematic formatting) or parameter choices (mismatched to the data) but can also be due to legitimately large/complex data! The first case is much more common when working with the clusters at the UseGalaxy servers. Examples → exceeds-memory-error.
From the part of the your message posted, we can’t see which data reports are being graphed but this is where we would suggest reviewing first. If these were the same reports that graphed interactively, then the basic format is likely Ok and the content is what to focus on. Is there anything special about this data? If you try to graph just 2 of the 12 reports with flat plots, what happens?
If you need more help, we can try to help more! Off hand, I’m guessing unusually dense data for the graph type. Maybe we can assist in modifying the job to avoid the excessive memory use.
Then, if these jobs seem like they should graph Ok but for some reason are not, we can help to review the technical details and possibly make adjustments to how the job is executing technically on our side. Learning which server you are working at would be part of this.
For this closer review: please capture more screenshots (the job details view’s input/parameter table with at least one expanded peek view of an input report plus all MultiQC parameters) or share the exact example error job in a shared history link (we love these details!). How to → How to get faster help with your question
Let’s start there and please let us know what happens! 