Monitoring Jobs

Another major feature of Oliver is the ease with which you can interrogate workflow status(es). In the simplest case, you can use the oliver status subcommand to get a summary of running jobs much like an HPC environment.

oliver status

To see individual workflows, you can use the -d/--detail-view parameter to show a "detailed" view.

oliver status -d

You also might be interested in a summary of the status of what steps each workflow in on. For this, you can use the -z/--steps-view parameter to show the "steps" view.

oliver status -z

By default, Oliver only shows all statuses. You can use the -r (Running), -s (Succeeded), -f (Failed), and -a (Aborted) flags to select a subset of these.

# only show failed and aborted jobs
oliver status -fa

Batching

If you run hundreds or thousands of workflows, Oliver's response time might become slow because it needs to retrieve information about every workflow Cromwell has ever run. In this case, it's typically useful to think of your job runs as self-contained groups of jobs separated by time called batches.

Oliver has facilities baked in to compute batches automatically (meaning you don't need to specify the batch number at submit time). Just ensure the batch_interval_mins configuration variable is set and try oliver batches.

oliver batches

You can restrict oliver status to only report only a subset of batches. For instance, you can report on the first batch ever run (-B/--batches-absolute).

oliver status -B 0

Or, more commonly, you are interested in the last batch run (-b/--batches-relative).

oliver status -b 0

Job Names and Groups

If you have submitted jobs with an Oliver Job Name or Oliver Job Group, you can restrict results to only workflows assigned to that name/group.

oliver status -g CohortName