This discussion has been locked. The information referenced herein may be inaccurate due to age, software updates, or external references.
You can no longer post new replies to this discussion. If you have a similar question you can start a new discussion in this forum.

Polling Intervals - Best Practices

FormerMember
FormerMember

Hello,

I had a tech reset all intervals back to default values.

The original Tech who setup the system has left the company

Is there a best practice to setting these polling intervals?

  • That's one of those "it depends" kind of questions. But the general rule of thumb is that the shorter the polling interval, the more strain on a polling engine (and more granularity in polling results). The opposite is true as well (more time between = less strain = less granularity)

    With a few exceptions, I have typically found that the default polling intervals are a great starting point for any environment. After 30 days, look at your historical data and you can then tell if you need to adjust based on granularity of data and the "load" on your polling engines.

  • FormerMember
    0 FormerMember in reply to zackm

    Thanks Zac,

    I figured the default settings would take a heavy hit on the processor engine.

    I wanted to get back to where it was when the original guy set it up.

    He was a genius and had this running perfectly.

    I'm sure that keeping it at default won't hurt anything more than generating a lot of alerts.

  • You can check you polling engine capacity by going to Settings > Polling Engines from the website.

    For each polling engine, there will be a 'Polling Rate', this will be a good representation of your server's capacity for polling at the current intervals. (There is a more detailed explanation of this metric under the 'Learn More' link.

    Polling Rate.jpg

  • I don't know about best practice, but here's how I use Orion.

    I use Orion to several needs, I would say:

    1. Normal Orion use, to see what's going on in network.

    2. To save historical traffic data to see trends and when network should be updated, which of course mean big DB (about 70 GB now).

    3. As alert system for device failures and real time monitoring, which mean frequent polling rate. Actually, even more frequent rate would be better, to show node status real time (as OpenView shows, which I used before).

    My settings may be different from typical user, but here are they in case you're interested:

    pollinginterval.png

    My polling engine looks like this:

    pollingengine.png

    So, as Zackm told, everything depends on what you want to see. Some may want to poll everything in every 10 seconds period, but doesn't need history more than few hours. Someone is interested in long period trends only...

  • Does anyone know about the best practice?
    Why most of the polling intervals have been set for 5 mins?

  • Default polling intervals here:

    SolarWinds NPM Online Help

    Polling Settings

    The Orion Polling Settings view establishes both the polling intervals that are observed by the SolarWinds NPM poller and the retention periods maintained for the SolarWinds Orion database. The following settings are configured directly from the Orion Web Console.

    Note: Changes to database maintenance and retention settings do not take effect until the SolarWinds Network Performance Monitor service is restarted.

  • Best practice will vary depending upon the number of elements you are polling, and how much strain monitored devices and your poller can handle. The more frequently you poll an interface for example, the more granular your statistical averages become. Some network devices can handle being polled more often than others. Keep an eye on CPU load of these devices and your poller rates. When you reach 80% it may be time to look into an additional poller.

  • With so many admins and upgrades and changes, it would be nice to have a button that resets everything back to what Solarwinds deems as the best practice settings.