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

Polling question

Hi,

We will soon be polling 10000 elements over a large WAN link (Orion one end and multiple sites at the other end), I have 2 concerns, hopefully someone can shed some light on them:

1.) When polling these elements from our NPM server and our additional polling server, are the polls (pings & SNMP stats collections) randomised? Or all polled at the same time?

2.) Should the WAN link have issues either our end or ISP etc we would get 1000's of emails saying devices are down, possible causing a DOS attach on our SNMP server, how can we safe guard against this?

Thanks

Find more posts tagged with

Accepted answers

All comments

bshopp

So couple things, what is the latency on average on the WAN link?

I am going to check into #1 further

#2 how often is something like this going to occur? I assume this downtime is not scheduled? So tell me more about the other sites. Are they all over the world or what? Might be worthwhile to look at NPM at each site and roll up to an EOC console, that way you don't have to worry about the main link going down.

bshopp

Regarding #1, they are basically randomized is the answer I got from Dev

Karlo.Zatylny

Hi,

Answer to #1.

Everything in the polling is based on a timer. The pings will happen by default every 2 minutes from when the node was added or the poller started. The statistics every 9/10 minutes. There is some drift so that a polling might be a second late or early depending on the busyness of the server. Also, there is some built in throttling to not blast a single device or network. You can see the with your maximum outstanding SNMP requests. With a busy poller, you will end up getting a fair amount of equal distribution as the polling continues.

We do not necessarily group all the polling for a single node and its interfaces together as each interface will have its own polling procedure executed. Thus the SNMP requests to a single node could be distributed throughout the entire interval or grouped anywhere along the way.

Thanks

GonzoUK

1 = Thanks

2 = latency from NPM to remote devices is 15ms to 25ms. The sites are all over London/UK, we don't ave the budget at this time to add to all these sites, something to think able moving forward.

3. When we plan and downtime on our WAN link or firewalls to stop the email alerts I simply disable the alerts, is this a good practice?

4. Also I haven't tried this yet but when I reboot the NPM server we get lots of fresh emails saying about all the disk space alerts on devices, say this servers c: drive in 95% and this servers is 99%, we don't want this, would disabling the alerts before rebooting solve this? This also happens when I upgrade to the next service pack, would disabling the alerts again help?

Many thanks

bshopp

Another idea is to unmanage the applicable nodes/interfaces

GonzoUK

Is there a quick way of doing this like the disable alert tick box?