GRID - Cluster/Overall Pending Jobs

This graph shows the number of batch jobs in a pending state over a particular time interval. One graph is produced for each cluster that is monitored by RTM. If the number of pending jobs is increasing, it indicates there is more workload than there is capacity. Jobs pend when resources of the specific type are needed but are not available on the cluster so it is possible to have jobs pending while the cluster is relatively idle. It is useful to look at this graph along with the pending reason-related graphs and dashboards in RTM to get insight into what resources are in short supply.

The following example shows a small cluster that was overwhelmed with a job array that contained over 40,000 elements. The chart shows the cluster clearly working through the 40,000 jobs over the process of days and drawing down the backlog.