Hello, working with one of our engineers for Avaya and they are convinced that the method Prognosis uses for SNMP causes issues with the PBX. It takes a performance hit, increases CPU, whatever. Is there a way to modify how Prognosis walks the PBX? Can we specify just specific OIDs instead of a complete walk? Can we alter the frequency?
Hi @MattAnderson ,
The snmp polling interval is defined in the solution files. Please log a support case so that our team can investigate the root cause and determine which interval will need to be adjusted. Please include irfax from monitoring server, the performance issue on the Avaya end in the case.
Depending on the solution, Prognosis walks either the whole MIB file or the particular OIDs, changing the polling from complete walk to a portion of the MIB will need to modify the solution which requires IR professional service team involvement. Once we find out the root cause of the problem we will be able to come up with the solution.
Has the cause of the performance issue been found?
I have not had the cycles to open a case on this. Are you experiencing similar issues?
we also had the discussion with Avaya whether it is allowed to install virus scanners and monitoring agents on the servers. A few years ago this was still a problem to get support in case of a failure. The virus scanner for the AACC servers (Windows OS) is installed and some directories are excluded from the scan. With the agents it is not so easy, on the Windows-based servers we have installed them. On the other servers (CM, SM, AES, etc.) or gateways SNMP users are configured, but we do not query from Prognosis, but from CheckMK . In so far we have a workaround there and two monitoring systems in use. Whereby CheckMK is more for the infrastructure (OS, Hardware) and Prognosis for the application telephony, there are also some traps that are sent for errors in the application, but these are unfortunately not displayed in the Prognosis. There we still have a blind spot.
Any trap sent to Prognosis can be alerted on. You just have to customize a threshold for this. Also, make sure the MIBs are loaded. I have examples if needed.
It would be nice if you could provide the examples. Unfortunately we are not very far along at this point.
Not sure your level of knowledge so apologies if not providing enough details. I'm assuming you have snmp enabled in the Prognosis config. I'd suggest first creating a display showing all of your traps coming in so you can see if you are receiving the traps you want to alert on using the snmptrapdetails data fields:
Create a threshold like below. The where clause doesn't require the regex unless you have standard naming conventions and want to only alert on specific devices. You can remove the last AND statement to keep it simple:
Check interval of 30 seconds since you are looking back 30 seconds: