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

Do I need another Polling Engine for NPM?

I am looking to solve two issues and wonder if this is the right direction. From a very high level view, we have a US and Europe site but under 10K devices we monitor. I am in the US. Currently, our primary and only polling engine is in Europe. In the US, we will get alerts that devices are down but when we check either Orion or the devices, they are up. This seems to be happening a lot. The second thing is that all the down alerts are being recorded in Europe time. So we are having to take 7 hours off the time reported that the device went down. My question is, would installing a polling engine help resolve these false alerts and would it fix the time issue? Would this also help reduce network traffic?

Find more posts tagged with

Accepted answers

All comments

dhanson

A Europe based polling engine could curb your false positives on node down, but so could adjustments to your poll intervals, or modifying your alert logic to compensate for possible poll failures in EU devices (ex. increase time before considering a device down). Note that you can filter this based on the device's name, or create a custom property to determine which node down alert to use for certain devices. For instance, a separate alert for EU devices that gives them 1 or 2 extra poll intervals before considering them down. It adds a minute or two to your reaction time in the case of a legitimate event, but could reduce the false positives, and save you more time in the long run. A common issue for this is timeouts...perhaps your polling has a timeout that is too high for some of the latency you irregularly experience in EU?

If the NTP source is in UTC, and the timezone is set to something like "clock timezone UTC -7", then it converts accordingly to set timestamps in traps and log messages from that device. We use UTC globally on all our devices, so no matter where an alert is coming from it appears in our system as UTC. If yours are set based on the timezone they are in, SolarWinds (I think) won't do any conversions for you, and will display the messages based on the device's timestamp. Perhaps

this is the problem?

HTH!

-Dan

singlebuoy

You're saying I can set different polling for different groups? We do have separate groups for US and Europe devices. The alerts are def being recorded in Europe time zone. (I am in the US) Im not sure about what NTP is set to and I wouldn't be able to change those settings anyways even if it was the issue.

dhanson

Under Manage Nodes, select a node and hit Edit Properties. Here, scroll down a little bit, and under "Polling" is an option for polling for status. You can edit this to change the poll interval for a specific device. Keep scrolling down and you'll see "Alerting Thresholds", and among them is response time. You can edit these to deviate from the global intervals and timeouts in the Settings menu. You can also perform this change to multiple nodes at the same time, so feel free to select all of your Europe nodes and make small adjustments to these.

With a custom property defining what's in EU and what's in the US, you can create alerts that recognize these as variables in the rule. With that, you can define two "Node Down" rules, and modify the Europe alert to trigger only if the condition has existed greater than a set amount of time. With a large enough wait time, you can elapse an additional polling interval or two to check status before it alerts.

HTH!

-Dan

pratikmehta003

I guess u need to re-look into your polling and threshold settings as mentioned by Dan as well. U can also check your 'node I don't see any reason for a new Polling engine here.

For the timezone it will take based on where your polling engine is setup.

singlebuoy

If the timezone will take based on where the polling engine is setup then I need another polling engine!

pratikmehta003

Just for the timezone issue, I don't think its a good idea to bring up one more polling engine.

Check with your Windows team here if they can help you. If u make any changes then even ur DB server should have the same time zone...

And where exactly you want the alerts to be shown with correct timezone? Only in Console or elsewhere?

singlebuoy

The whole setup is in Europe. DB and everything. We get email alerts in the US on devices but with European time stamps.

singlebuoy

We want the email alerts to show the device events in US time. The emails themselves are in US time, but not the email content which are the Orion alerts.

RichardLetts

The timezone is NOT based on where your polling engine is setup.

In general data flows:

device -> module engine -> database

alerting data flows:

database -> alert engine -> recipient

so the date/times you see in the emails from the alert engine come from its timezone

singlebuoy

So do I need to open a support request with Orion? Because I have no idea how to fix that.