Cisco CUCM decrease Inventory polling intervall only

[Steffen_Muessig]

Hi all

We see in our Cisco setup a CPU peak for every hour since Prognosis is connected.

Is there a polling intervall of 1h set for any data request on Prognosis?

If yes, can we change it somewhere ?

I assume it is the Soap API inventory readout.

The performance profiles Low to HIGH make no difference.

Many thanks 

Steffen

[SCOTT_BALDWIN]

Hello @Steffen_Muessig,

Prognosis uses many timers for the data requests it sends to CUCM servers. Changing the Prognosis Monitoring Profile for the cluster (as you mentioned in the post) is the first way to adjust the frequency of the queries sent by Prognosis to the CUCM servers.

The issue of CPU has happened a few times. The issue is most often associated with:

• The SOAP/AXL/Perfmon queries sent by Prognosis to the CUCM server which connects to the Tomcat (web server) process.
• Each time IR Support has checked into these issues, something in the environment is putting heavy pressure on the Tomcat process already and the combination of the existing use of Tomcat service plus the queries from Prognosis together overwhelms the Tomcat process on the CUCM server.
• Cisco Jabber clients registering via web is the one of the items we have repeatedly seen causing the overload of Tomcat.
• Often the Server Access Logs (from RTMT) need to be reviewed to know what is putting the cumulative pressure on the Tomcat service of the CUCM server.

Each environment is unique, so there multiple potential answers for this issue. It would be best to open an IR Support case so the actual cause is isolated.

Thank you,
Scott Baldwin