Hi
We are currently running OBSERVIUM Professional Ed with 31 Cisco 7600/3560/2960 polling around 5000+ ports.
The Traffic graphs are showing gaps as the poller seems to be taking a lot of time.
We have tried different values for the poller wrapper threads (16, 32, 48, 64) on a 16 core/64GB RAM machine but not seen any improvement.
We have also tried different values for the ‘ma-rep’ field for snmp bulkwalk (5, 10 etc) – had seen improvement initially after enabling with 5, but the polling times seem to have increased automatically after some time.
We are currently running with 64 Poller threads and max-rep 10.
Pls suggest if we are missing something.
A few screen-shots are attached.
Best regards
Gagandeep Mann