We have 5 pollers, 1 dedicated web server, and 1 dedicated SQL server. We have 13500 elements spread across the pollers. Also, one of the pollers is receiving Syslog, and another of the pollers collecting Netflow (though very little at this time). When our system was put together, I believe we went with what was recommended at the time. I don't have the hardware specifics, but for now, can we just assume it is at least adequate?
Our problem is that the performance of the Web Console is not satisfactory [for us]. In fact, it is often very slow. I would say that it is normal for any link we click in Orion NPM Web Console to take around 6-10 seconds minimum, and node details typically 15 - 20 seconds. But, they can often take even longer. If you put together a page with several charts of historical data, even worse.
I don't think this has anything to do with the pollers themselves. Why would poller performance affect web console performance? They are completely separate boxes performing different functions.
Can some of you tell me what kind of responsiveness you have with your web console compared to your setup? Are any of you monitoring larger networks, with little or no latency issues? I am trying to understand if we have a real problem, or if this is just typical performance for the Orion Web Console.
Thanks!
