To get some basic resource allocation by each container run, we should listen to the stream of docker stats, or call docker stats --no-stream periodically. In addition, the runtime of each tool should be monitored. If that works somehow, these logs should be added as metadata to the step tarball.
This will be a two-step issue: