0

I'm looking through the azure condition options.

One of the options is "Checks Failed Percent". It's an azure alert rule on a VM (Metric)

Can't find any data online, what will trigger the rule?

Let's say I have it like this: Whenever the average check failed percent is greater than 2%

What will trigger the rule? Is it when the CPU and Ram hit X % of CPU usage?

Anyone got a link, where you can read about all the conditions?

Peter Bons
  • 26,826
  • 4
  • 50
  • 74
Daniel
  • 97
  • 1
  • 8

1 Answers1

0

It is a metric emitted by the Network Monitor. The description reads "Percentage of failed checks for a test.".

Connection Monitor provides unified, end-to-end connection monitoring in Azure Network Watcher. The Connection Monitor feature supports hybrid and Azure cloud deployments. Network Watcher provides tools to monitor, diagnose, and view connectivity-related metrics for your Azure deployments.

Network Monitor lets you set up tests to test the connectivity between a VM / VMSS, for example

Your front-end web server virtual machine (VM) or virtual machine scale set(VMSS) communicates with a database server VM in a multi-tier application. You want to check network connectivity between the two VM/or scale sets.

The "Checks Failed Percent" condition allows you to create an alert when a number of tests have failed,indicating a loss of connectivity.

Peter Bons
  • 26,826
  • 4
  • 50
  • 74
  • Thanks for the answer. So it has nothing to do with CPU and Ram usage, it's only the loss of connectivity to the VM? To the people who have been working a lot with Azure, what % do you use in general? – Daniel Sep 14 '22 at 12:18
  • *So it has nothing to do with CPU and Ram usage* -> Correct! , *it's only the loss of connectivity to the VM?* Yes, but alerting on this metric only has sense when network monitor is configured, like shown [here](https://learn.microsoft.com/en-us/azure/network-watcher/connection-monitor#create-a-connection-monitor). Otherwise this metric will never have a value – Peter Bons Sep 14 '22 at 12:24
  • *what % do you use in general?* It depends. A connectivity issue might be transient so it is best to determine a base line based on previous measurements and use that as input for the value. – Peter Bons Sep 14 '22 at 12:25
  • There are no previous measurements, it's a brand new setup. But at 2% I see often the alerts are going off and being resolved 1-2min after. – Daniel Sep 14 '22 at 13:37
  • I'm just trying to figure out what 2% will trigger compared to 10% and I find it hard to find information on when each % will trigger. – Daniel Sep 14 '22 at 13:39