Hello and welcome back to Cyber Aires. Microsoft Azure Administrator A Z one of three course, I'm your instructor, Will Carlson. And this is Episode 11 about azure metrics.
In today's episode, we're gonna discuss what as your metrics are and where those come from, we're gonna discuss the properties of what every metric is goingto have, and we're gonna view some metrics in the azure portal.
So to get started, we're gonna jump here in the portal
to get started with metrics. We're gonna go in here to monitor into the monitor blade, and we're gonna talk about a number of these features as part of the easy 103 course. But our start here is gonna be with metrics and metrics in the azure ecosystem. Come from a number of sources, those air gonna be platform metrics, application metrics,
virtual machine custom metrics, and then custom metrics that you can send in via the rest. A p I the platform metrics are gonna be those that are created by azure. Resource is by default, such as our virtual machines have some metrics that are set up when you set up to be in
application metrics are gonna be those that come in from a custom application that you may have built If you've instrumented the application to send metrics toe azure when you built it,
you can also send in custom metrics from virtual machines using the windows diagnostic extension for Windows or the influx Data Telegraph agent for Lennox. And again, those custom metrics are gonna extend the metrics that air on by default for virtual machines
and last, like custom metrics of the arrest a p I if you encode it and send it envy arrest, you contract those metrics in azure as well.
No metrics are going to be collected every one minute by default, and they're gonna be retained for 93 days unless you send them off to a storage account or log analytics workspace.
Now, what we're looking at here is the metrics dashboard, and it is pretty blank. So let's collect Aries. Select a resource here.
We're gonna leave this in all subscriptions.
I'm gonna limit this to the I T resource group
and then I'm innocent. See, the resource is of this service desk could be in
Now, I'm gonna select my metric. Let's do
I'm here at the bottom
and I can see over the last little bit. There has been no CPU utilization for this particular
BM and I know that to be true because this PM has been powered off for a little while.
So what can I do about that? Well, I can change the time frame that this is looking at. Let's look back over the last seven dates
and we can see that the average CPU utilization of this B M has never been very high. But there definitely was a time when this CPU with the SVM was turned on and the CPU was being used. Now let's set another metric here.
We're gonna go up here to add metric
and let's do something
totally different and let's do
we're gonna do network in total here.
Now you can see that this completely blew up my graph. And why is that? Well, well, the difference has to do with the range of values put in here. So CPU utilization was only about 1% and
network in total was much higher than that 183 megabytes.
So that's blown the grap. I do that on purpose. To call out that you've got to be careful on the units of all of the measures that you put into a particular metrics.
So if I leave this initial CPU now, this makes a whole lot more sense. So what if you want to see multiples of these? Well, it's relatively simple. You just add a new chart and put in the metrics that you want to see on that particular chart so we could go back here and put back in our
percentage CPU. And now I've got both metrics up at a time. So something else it's powerful about metrics is that
you can pin these metrics to the dashboard. So if these air particularly relevant metrics that you want to see as part of a dashboard, maybe you wanna have a CPU dashboard for all of your V EMS so you can watch CPU utilization across all of those beings. You could create a dashboard and do just that
by setting up the metric here, making sure it looks like you want, and then pinning that to the dashboard of your choice.
We'll talk about alerts and a future episode, but you can also create alerts here from these metrics. So a good example real quickly would be if you wanted to be notified when CPU utilization spiked above 90%
you could create alert rule that would ultimately notify you that that CPU utilization head spiked. So really great to be able to do that here from metrics Because as you're exploring the metrics, you may find some things that you want to be alerted about that you didn't necessarily know before instead of having to back out of metrics and then remember what it was you were setting. When you're creating the alert rule,
you can create the alert rule right from metrics themselves.
Something else testable. And to know about metrics is that every metric and azure is going to contain four things. And that's gonna be a time stamp,
the type of miss metric,
the resource the metric is associated with and the value. And if we think through those really quickly, that completely makes sense. So for anything to be chargeable here, I've got to know when it happened, I got to know what it's measuring. So
this happened on July 7th at 11 53. It's measuring a virtual machine resource, specifically the service desk VM. That's the specific resource, and the value is going to be
2.66%. So again, metrics have time type, resource
something else to notice here. If I click off of this blade and go back to it,
Biometrics display is not saved.
Just a note. Don't spend a bunch of time building really pretty metrics Dashboard here and then expected to be there When it comes back, it will not. It will get deleted. If you like the metrics layout that you have sent him to your dashboard and look at them, they're on the ongoing.
One of the thing I want to mention about metrics is that they can be single or multi dimensional, and our CPU metric is going to be a single dimensional metric. It's got at a point in time. What resource and type is that? Metric and one value.
But there may be other metrics, particularly as it relates to a storage is a great example.
If we have a utilization of storage and we want to know out of that 80 megabytes, how much of that is used in each particular folder that would be a multi dimensional metric. It would all be rolled up into the one metric, and there will be multiple values split up for the folders within that metric.
And that's a little bit about metrics.
So in today's episode, we talked about the fact that metrics are collected every minute in azure. They have a time stamp type measurement, a resource that the metric was applied to and the value of that particular metric. So those four properties of metrics
we talked about the places that metrics originate, platform application, custom, virtual machine and custom metrics.
And we also talked about where you can go in portal to view those metrics. And that's the monitor blade on into metrics.
Up next, we're gonna talk about how you can collect even more details about the health of azure resource is that may not directly collect metrics by default.
Thanks for joining me today. I look forward to seeing you in the next episode