Just wanted to put a dedicated Feature Request for this one, related to the request for Fusion here

https://feedback.seqera.io/feature-requests/p/nextflow-fusion-needs-a-stage-in-method-to-optionally-copy-the-full-input-file-l

For the upcoming Seqera Intelligent Compute scheduler, it would be great if the scheduler could keep track of the NVMe disk usage in the EC2's which are running jobs, to ensure that new jobs dont get allocated to EC2 that might have a full or low free space in the NVMe. Not sure if this was already a noted feature or not. Thanks!