Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials

Inherited polling intervals - Node to interface

I ran into a node the other day that appeared to have data gaps when looking at the interface performance data at 5 minute intervals even though the node was set to collect statistics every 5 minutes. I started to question a long held assumption that I've had --

"Interface statistics collection rates are inherited from the node unless explicitly defined on the interface itself"

Any comments about my assumption? Maybe I just have a node issue and it is not collecting data regularly due to another problem -- or maybe my assumption has been wrong all this time,

Josh

Find more posts tagged with

SNMP

interfaces

Polling

Accepted answers

adatole

Then the Poller settings page under admin. The settings you care about (for purpose of this discussion)

Default Node Poll interval - this is ping.
Default interface poll interval - this is SNMP, but it only grabs the up/down status of the interface
Default volume poll interval - ditto for disks
Default rediscovery interval - this is how often SW will check "index numbers" to make sure interfaces and disks haven't re-numbered themselves. It does NOT go out and find new ones.
Default Node statistics poll interval - how often to get MOST statistics - CPU, RAM, custom OIDs, etc.
Default interface statistics poll interval - how often to get interface stats like bandwidth
Default volume statistics poll interval - ditto for disks

By default, your node, interface and volume up/down status is grabbed every 120 seconds.

...your node statistics every 10 minutes

...your interface stats every 9 minutes

...your volume stats every 15

Although in older versions those defaults were 5, 7 and 15, respectively.

The reason the node and interface stats are off by a minute is just to ensure the poller doesn't get bogged down pulling tons of stats all at once.

But the upshot is that, unless you have set them to be the same, you are likely NOT collecting your interface metrics at the same rate as your node metrics.

All comments

adatole

Then the Poller settings page under admin. The settings you care about (for purpose of this discussion)

Default Node Poll interval - this is ping.
Default interface poll interval - this is SNMP, but it only grabs the up/down status of the interface
Default volume poll interval - ditto for disks
Default rediscovery interval - this is how often SW will check "index numbers" to make sure interfaces and disks haven't re-numbered themselves. It does NOT go out and find new ones.
Default Node statistics poll interval - how often to get MOST statistics - CPU, RAM, custom OIDs, etc.
Default interface statistics poll interval - how often to get interface stats like bandwidth
Default volume statistics poll interval - ditto for disks

By default, your node, interface and volume up/down status is grabbed every 120 seconds.

...your node statistics every 10 minutes

...your interface stats every 9 minutes

...your volume stats every 15

Although in older versions those defaults were 5, 7 and 15, respectively.

The reason the node and interface stats are off by a minute is just to ensure the poller doesn't get bogged down pulling tons of stats all at once.

But the upshot is that, unless you have set them to be the same, you are likely NOT collecting your interface metrics at the same rate as your node metrics.

jbiggley

Crap.

So, now the question is whether or not there is an easy way to edit network interfaces to poll at a specific rate without editing them one-by-one? In this environment we poll based on an impact scale, so current node statistics rates range from 5 to 15 minutes.

I have custom attributes for impact at the interface level in the design, but I haven't rolled those out just yet.

Ideas?

adatole

Settings, Polling Settings

From there you adjust the timers and click "Re-Apply..." (polling intervals for the up/down, or polling statistics interval for the data collection)

That will change your polling settings sitewide.

adatole

I just re-read your question and realized I should add this:

If you go to Manage Nodes, you can select multiple nodes and change the:

Node status polling interval (ie: ping)
Statistics collection interval (ie: standard CPU, RAM, etc)
Topology report information (which I didn't talk about before and I'm not going to start now)

This means you can group your nodes together and edit their polling settings as a bunch (rather than one by one.

"But what about interfaces? you ask Not to worry

Under managed nodes, change "the "show" drop-down to interfaces, and then select the interfaces you ened. Then click "Edit properties

You'll see a screen where you can set the INTERFACE polling intervals for that group.

It's not a one-click-and-it's-all-fixed solution, but it beats the heck out of doing them one-by-one.

jbiggley

Awesome! I didn't even think about doing it that way for interfaces. I already handle the bulk edits for the nodes that way, hence my feature request to allow the ability to edit the node status, statistics and topology intervals separately as if they were a custom attribute instead of editing the entire block together. (I submitted it via support, but I think I'll post the details on the feature request forum and try to get it voted up.)

http://thwack.solarwinds.com/ideas/2409