5.0

Dynamic Thresholds
vCenter Operations Manager defines dynamic thresholds for every metric based on the current and historical
values of the metric. The normal range of values for a metric can differ on different days at different times
because of regular cycles of use and behavior. vCenter Operations Manager tracks these normal value cycles
and sets the dynamic thresholds accordingly. High metric values that are normal at one time might indicate
potential problems at other times. For example, high CPU use on Friday afternoons, when weekly reports are
generated, is normal. The same value on Sunday morning, when nobody is at the office, might indicate a
problem.
vCenter Operations Manager continuously adjusts the dynamic thresholds. The new incoming data allows
vCenter Operations Manager to better define what value is normal for a metric. The dynamic thresholds add
context to metrics that allows vCenter Operations Manager to distinguish between normal and abnormal
behavior.
Dynamic thresholds eliminate the need for the manual effort required to configure hard thresholds for
hundreds or thousands of metrics. More importantly, they are more accurate than hard thresholds. Dynamic
thresholds allow vCenter Operations Manager to detect deviations based on the actual normal behavior of an
object and not on an arbitrary set of limits.
The analytics algorithms take seven days to calculate the initial values for dynamic thresholds. Dynamic
thresholds appear as line segments under the bar graphs for use metrics on the Details page and on the
Scoreboard page. The length and the position of the dynamic threshold line segment depends on the calculated
normal values for the selected use metrics. Dynamic thresholds also appear as shaded gray areas of the use
metrics graphs on the All Metrics page.
Hard Thresholds
Unlike dynamic thresholds, hard thresholds are fixed values that you enter to define what is normal behaviour
for an object. These arbitrary values do not change over time unless you change them manually. You cannot
fix hard thresholds with vCenter Operations Manager.
Key Performance Indicators
vCenter Operations Manager defines attributes that are critical to the performance of an object as key
performance indicators (KPI). KPI are weighted more heavily in the calculations that determine the health of
an object. Graphs of KPI performance appear before other metrics in several areas of the product.
Probable Causes
If the condition of the currently selected object is degraded, you can view a list of probable causes for this
degradation on the Details page. vCenter Operations Manager analyzes the metrics behavior of all objects that
are related to the currently selected object to determine the probable reason for its health degradation.
The probable cause metrics might be about the object itself, about its child objects, or about other related objects.
They are the conditions, or symptoms, that are the first steps that led to the degraded condition of the object.
For example, an increase in CPU load by a virtual machine might lead to performance problems for other
virtual machines on the same ESX host, or for the ESX host itself.
Metric Concepts for vCenter Operations Manager Planning
Preparing to monitor your environment with vCenter Operations Manager involves some familiarity with
metrics that help you to identify a problem.
vCenter Server presents a use-based model of metrics. vCenter Operations Manager presents a demand-based
model of metrics. Some knowledge of the metrics that affect the data and graphs is useful to determine what
to do next in a workflow.
VMware vCenter Operations Manager Advanced Getting Started Guide
8 VMware, Inc.