The Workloads folder displays a cluster-wide view of workload metrics on two tabs: Weighted Scores and Base Metric Values.
Select Workloads from the Workload Management folder in the tree view. The workload scores report the weighted and base scores.
Each instance tracks a set of load metrics. Load scores and metrics are computed for each combination of instance and logical cluster, and are determined by applying the logical cluster’s load profile to an instance’s workload statistics. The result is an overall load score and a set of weighted scores that represent the relative impact of specific instance attributes.
The Weighted Score tab displays the load score and weighted metric values for each instance and logical cluster combination. If an instance is associated with two logical clusters, there are two entries for that instance in the details tab.
The Weighted scores tab includes:
Instance – name of the instance whose workload is represented.
Logical Cluster – name of the logical cluster associated with the instance.
Load Profile – the load profile assigned to the logical cluster.
Load score – a computed value representing the overall load on the instance. Compare this unitless number across instances as a means of comparing workloads.
User connections – the capacity of an instance to accept a new connection, based on resource availability.
CPU Busy – - a measurement of how busy the engines are, and provides the same information as sp_sysmon. Determines an instance’s capacity to accept additional work.
Run-queue length – the number of runnable tasks on a system. Run-queue length measures the processing backlog, and is a good indicator of relative response time.
IO load – measures outstanding asynchronous IOs, which indicates the relative IO saturation between instances.
Engine Deficit – measures the difference in online engines between instances. In a cluster where all the instances have the same number of engines there is no engine deficit. However, in a two-instance cluster where, for example, “instance1” has four engines and “instance2” has two engines, “instance1” has no engine deficit but “instance2” has a 50% deficit because it has half as many engines as “instance1.”
User – the weighted value of a metric you specified in the load profile.
Because each instance can be included in multiple logical
clusters, each instance has one set of metric values for each logical
cluster it belongs to.
The Base Metric Values tab displays all workload values for each instance in the cluster. Since each instance has only one set of values, regardless of how many logical clusters it participates in, one set of values is displayed for each instance.
Instance – name of the instance.
% User Connections – percentage of configured user connections in use.
% CPU busy – percentage of time the instance was busy performing work. This is a one minute, moving average taken from all engines on the system.
% Run queue length – base percentage of runnable tasks on a system. Run-queue length measures the processing backlog, and is a good indicator of relative response time
% IO load – base percentage of outstanding asynchronous IOs
% Engine deficit – the base percent difference in the number of online engines among instances in the cluster.
% User – percentage value you provide for the user metric specified in the load profile.
