104
Workload Balancing determines whether to recommend relocating a workload and whether a physical host
is suitable for a virtual-machine workload by evaluating:
• Whether a resource's critical threshold is met on the physical host
• (If the critical threshold is met) the importance assigned to a resource
Note:
To prevent data from appearing artificially high, Workload Balancing evaluates the daily averages for a
resource and smooths utilization spikes.
When evaluating utilization, Workload Balancing compares its daily average to four thresholds: low, medium,
high, and critical. After you specify (or accept the default) critical threshold, Workload Balancing sets the
other thresholds relative to the critical threshold on a pool.
Default Settings for Critical Thresholds
The default settings for the Critical Thresholds are as follows:
Setting Default
CPU Utilization 90%
Free Memory 51 MB
Network Read 26 MB per second
Network Write 26 MB per second
Disk Read 26 MB per second
Disk Write 26 MB per second
To prevent the pool master from becoming overloaded, Workload Balancing automatically sets the pool
master's Critical Thresholds at lower values.
To change the critical thresholds
1. In the Resources pane of XenCenter, select XenCenter > your-resource-pool.
2. In the Properties pane, click the WLB tab.
3. In the WLB tab, click Configure WLB.
4. In the left pane, select Critical Thresholds.
5. In Critical Thresholds page, accept or enter a new value in the Critical Thresholds boxes.
Workload Balancing uses these thresholds when making virtual-machine placement and pool-
optimization recommendations. Workload Balancing strives to keep resource utilization on a host below
the critical values set.
Tuning Metric Weightings
Balancing workloads occurs if a resource's utilization is significant enough to warrant or prevent relocating
a workload. For example, if you set memory as a Less Important factor in placement recommendations,
Workload Balancing may still recommend placing virtual machines you are relocating on a server with high-
memory utilization.
