Be honest, how many unread alert emails are in your subfolders in Outlook? We all set alerts for all kinds of things to ensure we’ll be aware if there is a problem in our environment. The problem is just that we either get too many alerts and we get numb or we don’t get enough. Want to know the solution? Watch this episode of Quick tip.
In Azure Monitor you can now set metric alerts with dynamic thresholds meaning that ML will look at the pattern of your resources and only send you an alert if it deviates from its norm. You can also then apply this automatically to all other resources deployed in the same subscription or resource group. I think this will be helpful for all of us as it will reduce the noise of alerts we receive. And so, we can rest assured that if we see an alert, we probably should pay attention. It also takes into consideration if your resource is below the expected metric which could also point towards a problem. Just think of the use case of where the server has always high metrics on Monday, but low on Friday. If you set the alert static on 75% CPU usage, then if it goes to 84% on Monday morning, you’ll get an alert even though it was expected. Dynamic thresholds will take care of this for you. Check out the documentation here: https://docs.microsoft.com/en-us/azure/azure-monitor/platform/alerts-dynamic-thresholds