Greetings!! I'm just starting our NPM deployment and I've got a couple Polling Engine Tuning questions for the community. We are monitoring 14k interfaces with a single poller, which we know goes against "best-practices." However, I had a hard enough problem getting managment to purchase this product, and I just couldn't get them to purchase a second poller right now. Hopefully next year!! And if anyone is interested, I am replacing Zenoss with NPM. That being said, I've changed some of the default settings. Just wondering what others are using, and looking for feedback on what I've done so far. * Default Interface Poll Inteval (I think this is only used to poll SNMP up/down status of an interface)* Is it only used for up/down interface status? * Anyway to turn it off? Out of 14k interfaces, I can't think of a whole lot of them that I need notified on. Maybe I'm crazy * Set the value to 1200 seconds (max it would allow) * Default Rediscovery Interval (I think this is used to discover interfaces that may have come up, IOS version, General SNMP information)* Set the value to 720 minutes (12 hours) * I'm thinking this kind of thing shouldn't be changing often enough to warrant less and I was contemplating 24 hours * I've also modified the Polls Per Second per the Polls Per Second Tuning* Max Node and Int Status Polls - 31 Per Second * Max Stats Collections - 124 Per Second In general, is there any easy way to tell if I'm not completing a polling interval or if I'm creating gaps any where, short of just trying to find one that has a gap? TIA for any insights!! -Russ

* Default Interface Poll Inteval (I think this is only used to poll SNMP up/down status of an interface) - correct.. * Is it only used for up/down interface status? - correct * Anyway to turn it off? Out of 14k interfaces, I can't think of a whole lot of them that I need notified on. Maybe I'm crazy yes.. you only get notified if you set up alerts for up/down status.. you can have your poll interval set at 120 seconds but only alert on devices you are interested in. I would create a custom node property for devices and alert of that group. * Set the value to 1200 seconds (max it would allow) * Default Rediscovery Interval (I think this is used to discover interfaces that may have come up, IOS version, General SNMP information) - correct* Set the value to 720 minutes (12 hours) - i left mine at 10mins and have no problems.. i have my server team that needs Volume information urgently so 10misn works for us.. * I'm thinking this kind of thing shouldn't be changing often enough to warrant less and I was contemplating 24 hours - this makes sense.. every environment is different.. * I've also modified the Polls Per Second per the Polls Per Second Tuning* Max Node and Int Status Polls - 31 Per Second * Max Stats Collections - 124 Per Secon

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
Store

Polling Engine Tuning

russuhte

Greetings!!

I'm just starting our NPM deployment and I've got a couple Polling Engine Tuning questions for the community. We are monitoring 14k interfaces with a single poller, which we know goes against "best-practices." However, I had a hard enough problem getting managment to purchase this product, and I just couldn't get them to purchase a second poller right now. Hopefully next year!! And if anyone is interested, I am replacing Zenoss with NPM.

That being said, I've changed some of the default settings. Just wondering what others are using, and looking for feedback on what I've done so far.

Default Interface Poll Inteval (I think this is only used to poll SNMP up/down status of an interface)
1. Is it only used for up/down interface status?
2. Anyway to turn it off? Out of 14k interfaces, I can't think of a whole lot of them that I need notified on. Maybe I'm crazy
3. Set the value to 1200 seconds (max it would allow)
Default Rediscovery Interval (I think this is used to discover interfaces that may have come up, IOS version, General SNMP information)
1. Set the value to 720 minutes (12 hours)
2. I'm thinking this kind of thing shouldn't be changing often enough to warrant less and I was contemplating 24 hours
I've also modified the Polls Per Second per the Polls Per Second Tuning
1. Max Node and Int Status Polls - 31 Per Second
2. Max Stats Collections - 124 Per Second

In general, is there any easy way to tell if I'm not completing a polling interval or if I'm creating gaps any where, short of just trying to find one that has a gap?

TIA for any insights!!

-Russ

Find more posts tagged with

Accepted answers

kulastone

Default Interface Poll Inteval (I think this is only used to poll SNMP up/down status of an interface) - correct..
1. Is it only used for up/down interface status? - correct
2. Anyway to turn it off? Out of 14k interfaces, I can't think of a whole lot of them that I need notified on. Maybe I'm crazy yes.. you only get notified if you set up alerts for up/down status.. you can have your poll interval set at 120 seconds but only alert on devices you are interested in. I would create a custom node property for devices and alert of that group.
3. Set the value to 1200 seconds (max it would allow)
Default Rediscovery Interval (I think this is used to discover interfaces that may have come up, IOS version, General SNMP information) - correct
1. Set the value to 720 minutes (12 hours) - i left mine at 10mins and have no problems.. i have my server team that needs Volume information urgently so 10misn works for us..
2. I'm thinking this kind of thing shouldn't be changing often enough to warrant less and I was contemplating 24 hours - this makes sense.. every environment is different..
I've also modified the Polls Per Second per the Polls Per Second Tuning
1. Max Node and Int Status Polls - 31 Per Second
2. Max Stats Collections - 124 Per Secon

All comments

lag

This 'might' help... I've found that the polling limitation for each poller is about 1000 SNMP polls per minute. If you can tune the system to be within those specs, you should be ok.

I've never had much luck with the PPS tuning.

What does your hardware platform look like?

russuhte

I do have some monies available to upgrade the hardware platform. It is currently running on a single dual core proc. I just put through the paperwork to get a second proc ordered. I have the option of adding additional memory as well. It currently has 4GB, but I'm really not sure that I'm seeing any memory pressures just yet.

-Russ

kulastone

Default Interface Poll Inteval (I think this is only used to poll SNMP up/down status of an interface) - correct..
1. Is it only used for up/down interface status? - correct
2. Anyway to turn it off? Out of 14k interfaces, I can't think of a whole lot of them that I need notified on. Maybe I'm crazy yes.. you only get notified if you set up alerts for up/down status.. you can have your poll interval set at 120 seconds but only alert on devices you are interested in. I would create a custom node property for devices and alert of that group.
3. Set the value to 1200 seconds (max it would allow)
Default Rediscovery Interval (I think this is used to discover interfaces that may have come up, IOS version, General SNMP information) - correct
1. Set the value to 720 minutes (12 hours) - i left mine at 10mins and have no problems.. i have my server team that needs Volume information urgently so 10misn works for us..
2. I'm thinking this kind of thing shouldn't be changing often enough to warrant less and I was contemplating 24 hours - this makes sense.. every environment is different..
I've also modified the Polls Per Second per the Polls Per Second Tuning
1. Max Node and Int Status Polls - 31 Per Second
2. Max Stats Collections - 124 Per Secon

russuhte

Very good information!! Thank you!!