Job failures because of time limit or memory limit can either use an automatic relaunch of the jobs with increased time and/or memory limit.
Other ideas that were pitched around were:
To assign memory and time requirements based on image resolution or size at build time.
To launch 5 or 10% of new assessor jobs for a project as a precursor and based on the results, launch rest of the jobs with an increased time and/or memory limit.
Job failures because of time limit or memory limit can either use an automatic relaunch of the jobs with increased time and/or memory limit.
Other ideas that were pitched around were:
To assign memory and time requirements based on image resolution or size at build time.
To launch 5 or 10% of new assessor jobs for a project as a precursor and based on the results, launch rest of the jobs with an increased time and/or memory limit.
Thoughts?