Calculating Node Availability
The Availability Calculation setting on the Polling Settings view provides a choice
between the following two methods for determining device availability.
Node Status
The default method is based upon the historical up or down status of the selected
node. The selected node is polled for status on the Default Node Poll Interval defined
on the Polling Settings view. For more information, see "Orion Platform Polling
Settings"
If the selected node responds to a ping within the default interval, the node is
considered up, and a value of 100 is recorded in the Response Time table of the
Orion Platform database. If the node does not respond to a ping within the default
interval, the node is considered down and a value of 0 is recorded in the Response
Time table of the Orion Platform database. To calculate node availability over a
selected time period, the sum of all Response Time table records for the selected
node over the selected time period is divided by the selected time period, providing
an average availability over the selected time period.
Percent Packet Loss
The second method is a more complicated calculation that effectively bases the
availability of a selected node on its packet loss percentage. As in the Node Status
method, the selected node is polled for status. If it responds within the Default Node
Poll Interval defined on the Polling Settings view, a value of 100 is averaged with the
previous 10 availability records. For more information, see "Orion Platform Polling
Settings."
The result of the Percent Packet Loss calculation is a sliding-window average. To
calculate node availability over a selected time period, the sum of all results in the
Response Time table for the selected node over the selected time period is divided
by the selected time period, providing an average availability over time.
Note: The Percent Packet Loss method introduces a historical dependency into each
availability node record. In general, it is best to leave calculations based on Node
Status unless you specifically need node availability based on packet loss.
204
Calculating Node Availability
To Next Page