Load indices measure the availability of dynamic resources on hosts. Dynamic resources are properties of a host that change as the load on that host changes, such as available memory and CPU utilization.
Load indices are not configurable; they simply reflect the current state of resources on your hosts given the current load they are managing. For example, as more workload units are assigned to and being run on host A, host A’s CPU utilization load index increases. A load index may increase or decrease as the host’s resources are put under more load. For example on host A, the total available memory decreases as the load increases.
Load indices are measured automatically at fixed time intervals. Each index is individually monitored and has its own update interval (from one of the shortest intervals at 15 seconds for status, to the longest interval at 120 seconds for available temporary space).
Viewing load indices for one host provides an excellent snapshot of how that host is performing at a specific moment. For troubleshooting purposes, you may want to track the load indices of one host over time (for example, an hour or a day).
Viewing load indices for all your hosts provides an overall snapshot of how your cluster is performing under its current load at a specific moment.