Hi @Kevin_Lourd, I am so sorry you’re having trouble.
I had a quick look at the metrics we have recorded for your n8n cloud instance accor-voc and it appears the instance regularly requests more memory than available to it. So it seems one (or more) of your workflow executions is simply too demanding for your current n8n cloud tier, possibly because it fetches too much data at once.
Perhaps you want to consider upgrading to a larger plan (you can find the limits for each tier here) or self-hosting n8n (in which case you can directly monitor the memory consumption and can provision additional resources as needed).