Once in a while I review the number of monitored elements in NPM and find myself retuning both our polling engines a job I truly dread. Not because it is a complex or an arduous task but because poller tuning isn't really an exact science, it is more like pinning the tail on the proverbial donkey.
Mostly this is because the 'max poll' calculation supplied in the admin guide doesn't really work for us; admittedly it’s a good starting point but it is only a small part of the considerations to be aware of when it comes to comprehensively tuning the polling engines. For example, other considerations might be Avg disk queue lengths on SQL server, deadlocks, network slowdowns, packet queues, server hardware performance etc.
So for the benefit of myself and others in this community I am creating this post to hopefully give users a chance to share info about home grown tips and tricks you use to tune your pollers.
Please feel free to add any tips or nuggets of info you use other than what is included in the admin guide; maybe you have you're own calculation for Polls Per Second tuning? Equally if you have any questions about Poller tuning you can submit it here such as; 'Do 'down' nodes slow the polling cycle adversely affecting the polling completion percentage etc?
I’m hoping that with the communities help we can take some of the mystery out of the black art of poller tuning
I will add some tips I use myself when tuning soon