I have been reporting WAN Availability based on the Availability column in the ResponseTime tables. Since the availability reporting was completed, I discovered that a new set of tables for Interface Availability were added with the last upgrade. Having compiled the necessary logic to tie Interface Availability to each interface, a trend has emerged.
When I compare the Node Availability to the corresponding WAN Interface Availability, the Interface Availability is almost always higher than the Node Availability.
QUESTION: If the Node is unavailable, wouldn't it stand to reason that the WAN Interface is also unavailable and therefore Interface Availability should always be <= to the Node Availability? Can it be that if the node is unavailable the interface availability is simply unreported. If so, I don't think this number can be trusted.
If Interface Availability only reflects the availability when the node is available, it seems you would subtract the Interface Unavailability from the corresponding Node Availability to get the “adjusted” Interface Availability number.
Node Availability + (Interface Availability – 1.00)
.99 + .99 -1.00 = .98 = 98%