Creating Dependencies to Stop Alerting

Hi,

I am trying to stop alerts coming in for nodes behind a switch. I created a group for the nodes behind the switch and increased the polling for those device as to not trigger before the switch...this works for node down alerts. I can reload the switch and not get any "node down" alerts. We are also monitoring these devices behind the switch for slow response time, so those alerts still come in.

Is there a way I can include these "slow response time" alerts in the dependency somehow so I don't get them when I reload the switch?

Thank you, Pat

Find more posts tagged with

Alerts

Accepted answers

gkjono

Hey Pat,

When a node is unavailable due to a dependency the status changes to "Unreachable". You should be able to modify the condition of your alert to exclude down and unreachable nodes similar to what I have below.

2018-08-27 12_23_32-Edit Alert - _Copy of High response time_.png

You may also want to set the alert so that the alert condition must exist for a certain amount of time before it alerts, if you run into issues where it's alerting before it realizes that the parent is down.

2018-08-27 12_28_02-Edit Alert - _Copy of High response time_.png

I actually have some free time this week if you want me to help you with this. I am a field engineer for Loop1 which provides professional services for SolarWinds products. Of course, there would be no cost if I helped you out on this. Feel free to DM me, or email using the link below.

-Greg Kjono

Loop1 Systems: SolarWinds Training and Professional Services

LinkedIN: Loop1 Systems
Facebook: Loop1 Systems
Twitter: @Loop1Systems

All comments

gkjono

Hey Pat,

2018-08-27 12_23_32-Edit Alert - _Copy of High response time_.png

2018-08-27 12_28_02-Edit Alert - _Copy of High response time_.png

-Greg Kjono

Loop1 Systems: SolarWinds Training and Professional Services

LinkedIN: Loop1 Systems
Facebook: Loop1 Systems
Twitter: @Loop1Systems

hpstech

Hi Greg -

I was wondering if the alert example above is a secondary alert, in addition to the alert that the Parent alert?

I assume that the 10 minute timeout is for secondary Node alerts for child objects, not for the primary alerts on the parent.

Thanks for your thoughts on this. I'm just looking for the best way to manage/automate Parent/Child group alerting.

gkjono

Hey Pat,

You could scope this either way. You'll likely run into situations where a node is a parent of some nodes, but a child to other nodes as well. Adding in the "unreachable" condition would just keep it from alerting in the circumstances where the node is a child, but shouldn't adversely affect it if it is a parent to other nodes.

I use the "Condition must exist for" option quite a bit to keep alerts from being too noisy. For example, if you have CPU spiking between 20% and 90%, it's probably not an issue. However, if that CPU is sustained at 90 - 100% for 15 - 20 minutes I'd like to know about it. It can also be used in these circumstances where a status of child might get updated before the status of parent. 10 minutes is probably overkill in this situation and could probably be cut back.