6.3

Table Of Contents
n
Datacenter, Cluster, Host, Datastore
Alert Definition Best Practices
As you create alert denitions for your environment, apply consistent best practices so that you optimize
alert behavior for your monitored objects.
Alert Definitions Naming and Description
The alert denition name is the short name that appears in the following places:
n
In data grids when alerts are generated
n
In outbound alert notications, including the email notications that are sent when outbound alerts and
notications are congured in your environment
Ensure that you provide an informative name that clearly states the reported problem. Your users can
evaluate alerts based on the alert denition name.
The alert denition description is the text that appears in the alert denition details and the outbound alerts.
Ensure that you provide a useful description that helps your users understand the problem that generated
the alert.
Wait and Cancel Cycle
The wait cycle seing helps you adjust for sensitivity in your environment. The wait cycle for the alert
denition goes into eect after the wait cycle for the symptom denition results in a triggered symptom. In
most alert denitions you congure the sensitivity at the symptom level and congure the wait cycle of alert
denition to 1. This conguration ensures that the alert is immediately generated after all of the symptoms
are triggered at the desired symptom sensitivity level.
The cancel cycle seing helps you adjust for sensitivity in your environment. The cancel cycle for the alert
denition goes into aect after the cancel cycle for the symptom denition results in a cancelled symptom.
In most denitions you congure the sensitivity at the symptom level and congure the cancel cycle of alert
denition to 1. This conguration ensures that the alert is immediately cancelled after all of the symptoms
conditions disappear after the desired symptom cancel cycle.
Create Alert Definitions to Generate the Fewest Alerts
You can control the size of your alert list and make it easier to manage. When an alert is about a general
problem that can be triggered on a large number of objects, congure its denition so that the alert is
generated on a higher level object in the hierarchy rather than on individual objects.
As you add symptoms to your alert denition, do not overcrowd a single alert denition with secondary
symptoms. Keep the combination of symptoms as simple and straightforward as possible.
You can also use a series of symptom denitions to describe incremental levels of concern. For example,
Volume nearing capacity limit might have a severity value of Warning while Volume reached capacity
limit might have a severity level of Critical. The rst symptom is not an immediate threat, but the second
one is an immediate threat. You can then include the Warning and Critical symptom denitions in a single
alert denition with an Any condition and set the alert criticality to be Symptom Based. These seings cause
the alert to be generated with the right criticality if either of the symptoms is triggered.
Avoid Overlapping and Gaps Between Alerts
Overlaps result in two or more alerts being generated for the same underlying condition. Gaps occur when
an unresolved alert with lower severity is canceled, but a related alert with a higher severity cannot be
triggered.
Chapter 3 Customizing How vRealize Operations Manager Monitors Your Environment
VMware, Inc. 47