3.11
Polling
As described in 1.4 - Acquistion
Stacks, AutoNOC features a very powerful multithreaded, multiprocessing operations and
data acquisition core that can poll many different devices simultaneously.3.11.1 PollEvery Variable
How AutoNOC polls data is defined and managed with the %POLLEVERY
variable. For more information on variables see 4.2 - Variables.
The %POLLEVERY variable is defined for each device
object by default. The default value is "60" which means poll every 60 seconds
as is shown in the following screenshot:

This variable can be defined at the device level,
component level, or at the template level. So, for instance, say you want the Ping
probe to be called every 5 minutes instead of every 60 seconds, modify the Ping
probe template and add a PollEvery definition to the template probe variable and
that variable will override the value defined for the device.
Please note that customizing polling intervals can
result in decreased optimization opportunities for AutoNOC's operation engine. One result
of this is that, in an effort to serve acquisition requests on a timely basis, AutoNOC may
end up polling some shared variables more than once. As an example, Total Traffic
and Traffic In each share some of the same SNMP queries, when these probes are
fired at the same time, AutoNOC optimizes away duplicate queries cleanly, quickly and
efficiently. If the probes fire at different times, AutoNOC may miss out on some of these
opportunities resulting in more queries.
3.11.2 Adaptive Polling
An interesting thing to note is that the %POLLEVERY
variable is evaluated by the AutoNOC OSP Command Prompt under the scope of the object that holds
the variable.
What is powerful about the polling variable being an
expression is that opportunities for adaptive polling are present. For instance, you can
create a polling variable like "if(%L<0,60,300)" which polls the device every
60 seconds if there is currently a problem, and every 300 seconds otherwise! |