Latest / Microsoft Cloud IT Pro Podcast / Episode 383 – Securing Azure: Monitoring and observing your Azure estate
Transcript
- 0:03Welcome to episode
- 0:05383
- 0:06of the Microsoft Cloud IT Pro podcast recorded
- 0:09live on August 23, 2024.
- 0:11This is a show about Microsoft 360 5
- 0:14and Azure from the perspective of IT pros
- 0:16and end users, where we discuss a topic
- 0:19or recent news and how it relates to
- 0:21you. In this episode, we continue our discussion
- 0:24on security
- 0:25as we transition from Microsoft 365 to Azure.
- 0:28We kick things off with Azure Security Logging
- 0:31and Auditing before moving into Azure Monitor for
- 0:34collecting, analyzing, and acting on telemetry data. We
- 0:37also explore how it can help you identify
- 0:40trends and anomalies
- 0:41to help with your threat detection.
- 0:46We got the Logitech MX keyboard, the MX
- 0:48mini. Are you enjoying it? I don't know.
- 0:50I have a problem. Sound very clicky clacky.
- 0:53It's not clicky. It is definitely not a
- 0:55clicky keyboard.
- 0:56It is better than
- 0:57the macOS keyboard except that it doesn't have
- 1:00my fingerprint reader. I do like the fingerprint
- 1:03reader on
- 1:04my Mac keyboard. Like, they need to just
- 1:06sell a standalone fingerprint reader for logging in
- 1:09to, like, desktops.
- 1:11I too wish they would get there. But
- 1:13it's backlit,
- 1:14but it's definitely not like a NuPhy or
- 1:17a Keychron or a Clicky Mechanical. It's yeah.
- 1:21I would say it's a little bit more
- 1:23tactile
- 1:24than, like, the macOS keyboard, Little bit more
- 1:27key travel,
- 1:28but not a mechanical keyboard. So if you
- 1:30like backlit
- 1:32macOS keyboards that are not, I would say
- 1:34better, and if you want space gray, they
- 1:36do not sell a space gray
- 1:39mini I don't want the number pad. They
- 1:42don't sell a space gray one without the
- 1:43number pad. They only sell white. They don't
- 1:45have a black one with the number pad?
- 1:46They have a black one with the number
- 1:48pad, not a black one without the number
- 1:50pad. Go figure. Baby will get one in
- 1:52September. Only pros use number pads. That's your
- 1:54problem. You you gotta upgrade. You gotta be
- 1:56more you gotta put more pro in the
- 1:59pro
- 2:00of Apple Pro. Joshua Sharfstein: What do I
- 2:01need a number pad for? Joshua Sharfstein: What
- 2:03don't you need a number pad for? It's
- 2:04there to like mishmash navigate.
- 2:06Turn it into like hot keys so that
- 2:09you can do, like, window management or something
- 2:11with it. When I hit 1, it goes
- 2:12to lower quarter. When I hit 9, it
- 2:14goes to the upper right quarter. No. That's
- 2:16what keyboard shortcuts are for.
- 2:18I'm telling you. Who needs extra keys when
- 2:20I can push 5 keys at the same
- 2:22time?
- 2:24Use the numbers
- 2:25as a
- 2:26macro pad, and you'll be okay. That's what
- 2:29my stream deck is for, and it gets
- 2:30in the way of my mouse. Oh my
- 2:32goodness. Okay. Today. Yes. Real quick before we
- 2:34do that. I should be recording the correct
- 2:36mic. Do I not sound like I'm on
- 2:38the right mic to you? No. You're not.
- 2:40Really? Yeah. I blame Teams. I have an
- 2:42issue with Teams, though we're not getting into
- 2:45the topic today.
- 2:47It has to do with its ability to
- 2:49select the right audio device even though the
- 2:51right audio device is selected. That's the Teams
- 2:53audio driver, Core Audio d. Horrible name because
- 2:56it's not actually Core Audio for, like, the
- 2:58system, but they named their audio driver Core
- 3:01Audio d for Teams, which is like,
- 3:03I don't know why you have a daemon
- 3:05with the same name as the other daemon
- 3:07that does the thing and Microsoft got a
- 3:09Microsoft. But let's just talk about other ways
- 3:11that Microsoft is gonna Microsoft. I don't know
- 3:13what other ways Microsoft will Microsoft. Loop 2.0
- 3:15is out. That still doesn't allow you to
- 3:17secure
- 3:18a workspace with any form of group whatsoever.
- 3:24You're asking the wrong questions.
- 3:27The question isn't, can I secure it with
- 3:28a group? It's, why would I ever wanna
- 3:30do that? And then once you get over
- 3:32the hump of,
- 3:34you can't, you'll find another way through. Yes.
- 3:36You know what my other way through is
- 3:38gonna be? I'm gonna go create an Azure
- 3:40Automation runbook that loops through all of my
- 3:43teams or my groups, my Microsoft 365 security
- 3:46groups. So when to look at the name,
- 3:48it's going to create a corresponding workspace
- 3:50and loop, it is then going to iterate
- 3:52through every member in my group, compare it
- 3:56with every member of my workspace,
- 3:59and then add or remove users appropriately
- 4:02running
- 4:03every 10 minutes in an Azure automation
- 4:06runbook. No. I'm not gonna do that. Super
- 4:08easy. Barely an inquiry. Workspaces should be securable
- 4:11by group. End of
- 4:13story.
- 4:14Especially given that there are now notes for
- 4:16Teams meetings. If I'm gonna do a Teams
- 4:18meeting and my meeting notes are going into
- 4:21a loop component,
- 4:23said loop component
- 4:241 should be able to be assigned to
- 4:26a workspace
- 4:27so that said meeting notes can be in
- 4:29the same workspace with all my other loop
- 4:32things for
- 4:33said client. It means the way channels work
- 4:36and they can aggregate meetings and you can
- 4:38have multiple meetings in a channel, but it's
- 4:40technically still all in the same team and
- 4:41still Yes. Joshua Sharfstein: still in the same
- 4:43channel? Wouldn't that make sense? Joshua Sharfstein: And
- 4:44then secured by the same group security group,
- 4:47yes. Joshua Sharfstein: You tried to tell them
- 4:48how to fix the problem, not that it
- 4:50was actually going to be fixed, but let
- 4:52let me know how much your Azure automation
- 4:53costs you to run every 10 minutes.
- 4:55I think I can get it in for
- 4:57free. I figured out the math ones. 500
- 5:00minutes, if it runs for 1 minute, I
- 5:02can run it how many times a day.
- 5:04We're gonna have to figure that out as
- 5:05Copilot.
- 5:06So have you no. You're Scott, I've been
- 5:08up since 3 AM teaching, and I warned
- 5:11you I'm gonna be all over the place.
- 5:12I keep trying to get you back into
- 5:14it. Thoroughly enjoying this. Back on topic. So
- 5:16back on topic, we chatted in the past
- 5:18about
- 5:20security, and we started getting into securing the
- 5:23modern workplace through
- 5:25the lens of
- 5:26identity and things that are available within
- 5:29intra ID. So we talked about things like
- 5:32security baselines for identity within
- 5:35intra and conditional access and some other stuff,
- 5:38and that kinda naturally leads into a conversation,
- 5:41which we had last time, about
- 5:44Microsoft 65 Office 365 workloads. Just because you're
- 5:48in that identity space, you're all tied in,
- 5:50you're already in SaaS
- 5:53land, and in that software as a service
- 5:54land. So you've got identity as a service,
- 5:57and you have software as a service things
- 6:00that exist out there
- 6:01like
- 6:02SharePoint Online and Exchange Online that are dependent
- 6:05on this identity as a service in the
- 6:07form of IntraID.
- 6:09The other thing that IntraID governs and becomes
- 6:12the identity store for is also Azure.
- 6:15So we should really talk about the Azure
- 6:18side of the conversation,
- 6:20which isn't going to be constrained so much
- 6:23to identity. Identity is a component about it.
- 6:25Right? So, like, something for conditional access, the
- 6:28way conditional access can apply to
- 6:32your
- 6:32SharePoint online tenancy. Conditional access can also apply
- 6:36to things like the Azure portal or potentially
- 6:38workloads
- 6:39that you stand up, say, like a custom
- 6:40website that you secure with entry ID and
- 6:43you're driving, like, OAuth authentication through there,
- 6:46then that's available to you as well. So
- 6:49if you're interested in, like, the ID side
- 6:50of it, you can go back and listen
- 6:52to that episode. But for this one, I
- 6:54wanted to just talk more about
- 6:56Azure
- 6:57Security in general, which starts to get a
- 6:59little bit weird because security is an open
- 7:01ended conversation anyway. And then it gets even
- 7:04weirder because
- 7:05it's Azure, and what is Azure, but a
- 7:09set of components
- 7:10that all sit behind this governance scheme of
- 7:13things like subscriptions and management groups and ultimately
- 7:16that identity construct in IntraID.
- 7:19So, yeah, we should talk a little bit
- 7:20about Azure and some of the other things
- 7:22that sit over here. So things like security
- 7:25we talked about security baselines for identity. We
- 7:28should talk about security baselines for Azure and
- 7:30what those mean, and how as we start
- 7:32to decompose out of identity as a service
- 7:35and software as a service, and we get
- 7:36more into, like, platform as a service, infrastructure
- 7:39as a service components that are out there,
- 7:41what do we do, and how do we
- 7:42think about that? Yeah. And real quick, because
- 7:44you mentioned identity too, it made me think
- 7:46of this. Just a heads up with identity,
- 7:49there is also a new
- 7:52identity in Azure. Starting in October,
- 7:54so, like, a month from when you'll probably
- 7:57hear this episode if you listen to it
- 7:58on release day, Microsoft is actually making MFA
- 8:02mandatory
- 8:03for all
- 8:04Azure users. Your break glass accounts.
- 8:07Come up with a plan for your break
- 8:09glass accounts because they are going to have
- 8:12to be MFA'd.
- 8:14A great way to think about that
- 8:16is
- 8:16our friend,
- 8:18mister
- 8:19Yubike or a Fido key. Since you do
- 8:21have to have MFA,
- 8:23tying that to
- 8:25my admin today's phone number might not be
- 8:27the best way to go about that. Consider
- 8:30things like Fido keys and plan for that
- 8:32because now you're going to have that for
- 8:33your break glass accounts as well. That I
- 8:35think that's probably
- 8:36more the automation
- 8:38and how am I getting in on the
- 8:40back end if something does go wrong scenario.
- 8:42Yep. Because this is also then gonna roll
- 8:44out early 2025,
- 8:46gradual enforcement of MFA
- 8:48for Azure CLI, Azure PowerShell, Azure Mobile app,
- 8:51infrastructure as code tools, like, everywhere. And with
- 8:54the YubiKeys,
- 8:56like, I would say
- 8:57see, this gets a little tricky because you
- 9:00also don't wanna just hide to 1 YubiKey
- 9:01because have you ever misplaced a YubiKey and
- 9:04then you can't log in? It's let's do
- 9:06this for multiple YubiKey
- 9:08and probably have multiple administrators
- 9:11of your Azure environment have these YubiKeys. Because
- 9:14I'm with you. I've had it tied to
- 9:15a phone or tied to my authenticator app
- 9:17in my phone and then you lose your
- 9:19phone or you reset your phone. That MFA
- 9:21does not come back reliably
- 9:24in my experience
- 9:25with the authenticator app. I don't know if
- 9:28you start switching to passkeys.
- 9:31Yeah. I like YubiKey for it. Phishing resistant,
- 9:33secure. But that is one thing I did
- 9:35want to mention that people should absolutely start
- 9:38preparing for kind of in that whole Azure
- 9:41security and identity vein that has been announced.
- 9:44To your point of dual enrollment,
- 9:46even with a single key, you might wanna
- 9:48consider dual enrollment depending on, like, the devices
- 9:51that you might need to actually use to
- 9:52access.
- 9:53For example, I have
- 9:56a USB c UB key with NFC on
- 9:59it. I've enrolled both the NFC component and
- 10:02the USB c component,
- 10:03and that lets me use USB c, like,
- 10:05on my desktop or my laptop when I
- 10:07need to get in. But if my desktop
- 10:09or laptop aren't available and I still have
- 10:11the key on me because it's on my
- 10:12Keyring,
- 10:13then I can still get in via my
- 10:15phone, which does support NFC. And it has
- 10:18to listen. There's the call out to multiple
- 10:20keys also to minimum every time.
- 10:23Any piece of hardware can die. It it
- 10:25can crash out. It can crap out. I've
- 10:27had the YubiKey, particularly like the nano YubiKeys,
- 10:30the ones that stick in, the little nubbin.
- 10:32Those things seem to die on me all
- 10:34the time, and I don't know if it's
- 10:35just cause they get so tight in the
- 10:37USB c ports and something goes wonky when
- 10:39you're trying to pull them out.
- 10:40Whatever it is, like, those things have not
- 10:42been reliable for me. So I've completely moved
- 10:45over to, like, USB c plus NFC kinds
- 10:47of things so that I can make my
- 10:49life just a little bit easier. Also,
- 10:52consider that if you're doing FIDO keys, they're
- 10:54going to come in various forms of connectivity,
- 10:57not just USB C and NFC, but also
- 11:00USB A. And what's the device I need
- 11:03to log in on? So you might actually
- 11:04want to have 2 or 3 of these,
- 11:06especially in the cases of, like, break glass
- 11:08accounts
- 11:09because it is super important kind of stuff.
- 11:11Like, it's worth investing $30 in a couple
- 11:13of keys and then distributing them out there
- 11:15and getting them to where they need to
- 11:16be. 100%. I'm with you. I have 3
- 11:18keys
- 11:19that I enroll in almost all of my
- 11:21services when I use FIDO keys, and they're
- 11:23for sure all enrolled in Azure AD. Same
- 11:25thing as you. One's USB a, couple are
- 11:27USB c, one of them has NFC as
- 11:30well.
- 11:31Multiple options, they're not all in the same
- 11:33spot. So if I lose one, I can
- 11:35go look for 1 in another spot and
- 11:37all that. So moving on from identity and
- 11:40security to where you started to take me
- 11:43is
- 11:44security all up in Azure. Where do you
- 11:46wanna start with? Is this a broad, big,
- 11:48large topic? I think it's
- 11:51helpful
- 11:53to start at the top,
- 11:55consider what's the landscape of things that's available
- 11:58to you. So I I think one of
- 12:00the reasons this topic becomes
- 12:02so broad
- 12:03is you could say,
- 12:06I don't know, something like I'm a Azure
- 12:09SQL customer. Yep. And I'm running
- 12:12SQL
- 12:13inside of
- 12:15Azure
- 12:16Virtual Machines. I'm not running like SQL as
- 12:18a PaaS service, I'm running it as IaaS
- 12:20in a virtual machine. And in that world,
- 12:23you're running virtual machines. And how do you
- 12:25secure virtual machines, and what does that look
- 12:27like? And that's a pretty constrained conversation.
- 12:29Maybe that takes us back to the CrowdStrike
- 12:31thing. But if you take a step back
- 12:32from that and you say, hold on, like,
- 12:34before I was a virtual machine customer,
- 12:37what was I? Oh, I was an Azure
- 12:39customer.
- 12:40So what are the things that are available
- 12:42to me in Azure that I should go
- 12:45think about lighting up and turning on and
- 12:48enabling
- 12:49based on my scenarios?
- 12:51So in my mind, there's no reason that
- 12:54you would treat
- 12:55Azure in any
- 12:57way differently than you might like your
- 13:00Microsoft 365 environment. Right? So when you go
- 13:02into your Microsoft 365 environment,
- 13:05you don't first go into
- 13:07Exchange Online
- 13:09and start configuring security around Exchange Online.
- 13:12You start at the top and you say,
- 13:13oh, I'm I'm an m 65 customer. What
- 13:16can I turn on that gives me logging
- 13:17across the suite? And,
- 13:19hey, there's the audit log. So let let
- 13:21me go start to light that stuff up.
- 13:23What can I do to get visibility into
- 13:26the wider swath? And I think that's a
- 13:28good way to start with Azure as well
- 13:30because
- 13:31we've tied into the identity piece. So we
- 13:34talked about that and getting in, you can
- 13:36report on your sign in logs and start
- 13:38to understand that stuff. And then you take
- 13:40that next stop down to Azure,
- 13:42and before you even get into the services,
- 13:45there's
- 13:46this common base layer for Azure customers in
- 13:50the form of Azure Resource Manager and like
- 13:52the
- 13:54modern API surface, which is different than that
- 13:56old Azure service management service. At the ARM
- 13:59layer, you have these things like activity logs
- 14:04that are available automatically. Right? So for every
- 14:07Azure resource that's deployed,
- 14:09that you interact with, in certain cases that
- 14:12you perform like listing operations with or things
- 14:15like that, you've got the activity log there
- 14:18by default.
- 14:20And the activity log is integrated directly into
- 14:22a service,
- 14:24called Azure Monitor. And what does Azure Monitor
- 14:27give you? Azure Monitor gives you insights into
- 14:30metrics about your environment. So metrics being numbers.
- 14:34Right? Number of API calls, number of errors,
- 14:37number of sign in attempts, like whatever it
- 14:38happens to be given the the service you
- 14:41consume. So we start at the Arm layer,
- 14:43we start at the controllable plane. I think
- 14:45the first place we look is
- 14:48activity logs in general,
- 14:50because as you're getting into Azure, like you
- 14:53light it up and you have no resources,
- 14:56the very first resource you create let's say
- 14:59you create a virtual machine. So you're gonna
- 15:01go into the marketplace, and you're gonna click,
- 15:04I wanna create a new resource, you're gonna
- 15:06search for virtual machines, and you're gonna hit
- 15:08that blade. Ultimately, you're gonna submit that deployment,
- 15:11and that deployment, the act of submitting that
- 15:13deployment and having it ingested by Arm, captures
- 15:16that entry in the activity log, and then
- 15:18it's there for you. And then later when
- 15:20you come back and you interact with that
- 15:21virtual machine, let's say you come back and
- 15:23you add a disk to it, you're modifying
- 15:26that deployment,
- 15:27then that's captured in the activity log for
- 15:29you. And like I said, there's a whole
- 15:31slew of information that's captured in the activity
- 15:34log in general, but it's there for 90
- 15:37days, 93 days by default.
- 15:40It's basically 3 months of data, both Azure
- 15:42Monitor for metrics and your activity log as
- 15:45well that you just have out of the
- 15:48box,
- 15:49ready to go. You don't have to turn
- 15:50anything on. You don't have to configure anything,
- 15:52but you do have to know that it's
- 15:54there and that it's available to you because
- 15:57what's going to happen is, for that activity
- 15:59log, any interactions with the control plane, so
- 16:02things that happen through the API surface of
- 16:05Azure Resource Manager, that would be your CRUD
- 16:08operations for these resources, your creates, your reads,
- 16:11your updates, your deletes. All that stuff is
- 16:14going into the
- 16:16Azure Activity Log,
- 16:18and it's just there,
- 16:20ready to go for you, which is nice.
- 16:21You didn't even have to turn it on.
- 16:23It's always weird to me that you have
- 16:25to turn on the audit log in Office
- 16:26in M365,
- 16:28but I get why they do it, but
- 16:29it's well, just turn it on by default.
- 16:31It's Right. They had a security incident. Right?
- 16:33That compelled them. That made them turn it
- 16:35on, but you're right. For a long time,
- 16:38multiple years, it was not on by default.
- 16:39And I said the same thing. I'm like,
- 16:41that's stupid. It should just be on. Anyways,
- 16:43yes, it is. It's on. It's there. Use
- 16:45step 1 is recognize
- 16:47the activity log is there. One of the
- 16:50things that I mentioned was the activity log
- 16:53is
- 16:53integrated
- 16:55with
- 16:55Azure Monitor.
- 16:57Azure Monitor is this whole
- 17:00Azure service
- 17:01that is
- 17:03a crosscut, and it allows ingestion from a
- 17:05bunch of other Azure services.
- 17:07But the goal is to give you
- 17:10access to that observability data. That comes in
- 17:13the form of the activity log, which is
- 17:16effectively strings, right? We're talking about like words
- 17:19that go in, this was a create event,
- 17:21here was the name of the resource,
- 17:23this was the
- 17:25UPN
- 17:26that initiated that call, so you can do
- 17:28things like track, k. Was this resource created
- 17:31by,
- 17:32a service principal? Was it created by a
- 17:33real human? Which human was it created by?
- 17:35And how does that start to tie back
- 17:37to your identity environment? That's all there for
- 17:38you. And the other thing that it gives
- 17:40you is the observability
- 17:42part of the platform
- 17:44is it also gives you access to
- 17:47metrics.
- 17:48So you're probably, like, once you start deploying
- 17:50Azure services,
- 17:52and and it's not directly related to security,
- 17:55but it's important to know that it's there,
- 17:57is you're probably gonna wanna think about ways
- 17:59that you can tie metrics into that
- 18:01and things like metric alerts. Let's say I
- 18:04deploy
- 18:05a storage account. So I deploy a storage
- 18:07account, one of the things that I might
- 18:09want to monitor as
- 18:12a storage customer
- 18:14is
- 18:15how many transactions or how many errors am
- 18:17I driving? Like, how many transactions are erring
- 18:19out? What's the class of the error? Is
- 18:21it is it a 5 0X error, like
- 18:23a a throttle? Is it a 4 0X?
- 18:26Maybe it's 4 0 threes or 4 0
- 18:28fours, things like that. So that's all tracked
- 18:31in monitor by default. You can just go
- 18:33into monitor,
- 18:34and you can select a storage account, or
- 18:37you can even graph multiple storage accounts if
- 18:39you want to. I think you can put
- 18:40up to 200 individual storage accounts in a
- 18:43single Azure Monitor workbook for crosscut reporting,
- 18:46and you can immediately get those insights in
- 18:49a visual form.
- 18:50Or you can go the next click stop
- 18:52and you can say, hey, here's the things
- 18:53that are really important to me. Like, I
- 18:54might wanna know when I have more than
- 18:56n
- 18:57five zero three errors
- 18:59in a given period. So in the last
- 19:01hour, if I had more than a 1,050
- 19:04threes, maybe I've got some segment of throttling
- 19:07going on that's a little bit higher than
- 19:08I want it to be. So you can
- 19:10do things like configure alerts on top of
- 19:13that metric or those set of metrics, or
- 19:15you can start to combine the logs and
- 19:16metrics inside of alerts as well,
- 19:18and and you can start to
- 19:20layer that information
- 19:22together
- 19:23a little bit along the way, if that
- 19:25makes sense. Some people are like, yeah, platform
- 19:27metrics, my storage account, my
- 19:29CPU resources, that type of stuff. Maybe not
- 19:32security, but I also think as you start
- 19:34developing a baseline for how these should behave,
- 19:37a storage account or
- 19:38a VM that you're monitoring these platform metrics
- 19:42on, and you set those alerts to alert
- 19:45you when you stray outside of your baseline,
- 19:46it could indicate a security event. Like, someone
- 19:49got access to your storage account and started
- 19:50uploading and downloading a whole bunch of junk
- 19:52to it or somebody got into a VM
- 19:54and is running a bunch of processes on
- 19:56it that are and you see it because
- 19:58your VM CPU
- 20:00spiked and is running way higher than it
- 20:02normally runs. The SQL resource or SQL platform
- 20:05metrics, all of those things, I think, are
- 20:08very valid, and it's something that people would
- 20:10monitor
- 20:11anyways on prem. If you're running a VMware
- 20:15host or some other
- 20:17servers on prem, you're typically monitoring
- 20:20CPU consumption, memory consumption,
- 20:22storage,
- 20:23disk, IO, all that type of stuff. So
- 20:26why would you
- 20:28stop monitoring it when you go to Azure?
- 20:30Because,
- 20:31ultimately,
- 20:32Microsoft doesn't care if you fill up a
- 20:34storage account or if you incur a bunch
- 20:36of costs or have some type of security
- 20:38incident that's really affecting
- 20:40your data. Right? It's still that shared responsibility
- 20:44model of I'm running a VM in Azure.
- 20:46I'm still responsible for patching my OS. I'm
- 20:49still responsible for
- 20:50antivirus on the v maybe even storage accounts,
- 20:53like, getting into some of the other security
- 20:55options maybe that we'll talk about is
- 20:57what files are in my storage account that
- 21:00I have up in Azure. So I think
- 21:03from that perspective,
- 21:04like, platform metrics
- 21:06can still very much be an indicator of
- 21:09security issues or other issues. I I have
- 21:11a lot of conversations
- 21:12with customers in this area,
- 21:16is there's both the control plane and there's
- 21:18the data plane. So there's interactions
- 21:20that are driven through
- 21:22Azure Resource Manager, like I want to in
- 21:25the case of a storage account, I wanna
- 21:27create a storage account. Now once your storage
- 21:29account is created and you do things like
- 21:31you
- 21:32stand up a container and you upload an
- 21:34object into that container,
- 21:36the upload of that object, the creation of
- 21:38that container, those are data plane operations. So
- 21:41we've we've actually crossed over to a new
- 21:43API
- 21:44surface. But the the
- 21:46the thing to remember here is,
- 21:48like, in the context of, say, a storage
- 21:50account and just about every service I can
- 21:52think of off the top of my head
- 21:54that's in Azure,
- 21:55even though it's got a data plane to
- 21:57it, that data plane information was still available
- 21:59and monitored to me
- 22:01alongside my control plane information. So it's a
- 22:05kinda have your cake, you need it too
- 22:06scenario, and it's important to recognize, like, where
- 22:09those things are coming in and how they're
- 22:11getting there and where they're coming from. Because
- 22:13in that scenario where I described
- 22:16monitoring a storage account for things like number
- 22:19of errors,
- 22:20those errors, those those HTTP error codes like
- 22:23that 404,
- 22:24403,
- 22:25those 5 0 x errors, those are ultimately
- 22:27data plane metrics. But I didn't have to
- 22:29do anything to turn them on. Like, they
- 22:31were just there for me out of the
- 22:32box, and they were presented to me through
- 22:34that observability
- 22:35layer in Azure Monitor
- 22:37where it was automatically combining
- 22:40my control plane and data plane. So while
- 22:43Azure Monitor itself is not a security service,
- 22:47it has these components
- 22:49that are giving you all the observability and
- 22:51alerting capabilities,
- 22:53which
- 22:53leverage the right way,
- 22:55becomes something that's a component of your strategy
- 22:59to thinking about how you monitor and secure
- 23:02your environment. Because a lot of security constructs
- 23:04do come down to things like observability. If
- 23:07you think of even like the basics of
- 23:09an AV client, an AV client is there
- 23:11to watch for heuristics and new things to
- 23:14load and try and block them when they're
- 23:15malicious. So it's own kinda like little observability
- 23:18container. And, yeah, once you start to
- 23:21walk down that path and and recognize that
- 23:23crosscut is there for you, I I think
- 23:26it it makes things a little bit easier,
- 23:28and it opens up that mindset
- 23:31of, how do I get out of just
- 23:33observability
- 23:34mode into
- 23:36translating my security needs into things that observability
- 23:39is going to inform and where does this
- 23:41pull me to?
- 23:46Do you feel overwhelmed by trying to manage
- 23:48your Office 365 environment? Are you facing unexpected
- 23:51issues that disrupt your company's productivity? Intelligent is
- 23:54here to help. Much like you take your
- 23:56car to the mechanic that has specialized knowledge
- 23:58on how to best keep your car running,
- 24:00Intelligent helps you with your Microsoft cloud environment
- 24:03because that's their expertise.
- 24:05IntelliJunk keeps up with the latest updates in
- 24:08the Microsoft cloud to help keep your business
- 24:10running smoothly and ahead of the curve. Whether
- 24:12you are a small organization with just a
- 24:14few users up to an organization of several
- 24:161000 employees,
- 24:18they want to partner with you to implement
- 24:20and administer your Microsoft Cloud technology.
- 24:23Visit them at inteliginc.com/podcast.
- 24:27That's intelliginkdot
- 24:32com/podcast
- 24:34for more information or to schedule a 30
- 24:36minute call to get started with them today.
- 24:39Remember, Intelligink focuses on the Microsoft cloud so
- 24:42you can focus on your business.
- 24:46So that covers activity logs, platform metrics, those
- 24:50both collected by default have access to. There's
- 24:53one other type of kinda core
- 24:57log. I would say you probably also want
- 24:59to monitor high I would highly recommend you
- 25:02start monitoring, but this one is not on
- 25:04activity logs is I'm going in and I'm
- 25:07creating storage accounts, I'm creating VMs, platform, the
- 25:10metrics we just talked about. Resource logs are
- 25:13actually then
- 25:14insights into operations that are performed
- 25:17within the Azure resource. So actions that are
- 25:19performed within a
- 25:21storage account,
- 25:23uploading, downloading
- 25:25files,
- 25:26actions within a VM. You can pull out
- 25:28event logs from VMs,
- 25:30actions within the SQL database.
- 25:32You have to turn on your diagnostic settings
- 25:35in these individual resources
- 25:38or potentially use something like Azure policy
- 25:42or something else to turn these on and
- 25:44then figure out where do I want these
- 25:46logs to go. And these are going to
- 25:49very much vary
- 25:50resource to resource
- 25:52based on what types of logs you're collecting.
- 25:55Logs within
- 25:56IIS are very different than logs within storage,
- 25:58are very different than logs within
- 26:00VMs.
- 26:01You can pick and choose too. There's multiple
- 26:03logs within all of these. You can go
- 26:05in and say, I just want all the
- 26:06logs from this resource. Or based on the
- 26:09resource, you can go in and pick, I
- 26:10just want this certain type of information, these
- 26:12certain logs from this particular resource. And usually,
- 26:15you have different options. You can go into
- 26:17storage accounts with these. Scott would love it
- 26:19if you put all your logs in his
- 26:21storage accounts. Typically, I think what people tend
- 26:24to do, and this is
- 26:25what I tend to do, is send these
- 26:27into
- 26:28log analytics. You can also I can't remember.
- 26:31What are the are there 3 or 4
- 26:32options? Log analytics, a storage account, queues?
- 26:35There's 3. Log analytics, storage account, and event
- 26:39grid. Yes. So you can send them back
- 26:40through through eventing.
- 26:42Taking a step back, when we talk about
- 26:45monitoring and observability
- 26:46and then tying observability
- 26:48back to things like what do I want
- 26:50to monitor to
- 26:51think about the security of my environment,
- 26:54You need to start answering those questions
- 26:56like service by service or area by area.
- 26:59So for a storage account, I could be
- 27:02really interested, like I said, in things like
- 27:05number of errors
- 27:07on a given subset of transactions that exist
- 27:10out there. So that's my class of, hey,
- 27:13what are my 4 0 x errors for
- 27:15GitLab requests? Like, how many unknown requests am
- 27:17I getting? Is somebody randomly hitting me? Once
- 27:19you start getting to the place where you're
- 27:21like, show me the user who did that
- 27:24thing, or show me the URI
- 27:26for the resource that was impacted by that
- 27:29event,
- 27:30those are logs. So a good way that
- 27:32you can rationalize this in your head, or
- 27:34at least it's a way that I found
- 27:35to work for me, is metrics are always
- 27:38numbers,
- 27:38and those are always gonna be in Azure
- 27:40Monitor, and they're always gonna be there for
- 27:41free for the those 3 months. Outside of
- 27:44that, it's on you to retain them for
- 27:45longer. Once you start talking about things like,
- 27:48oh, was it a UPN?
- 27:50Was it a URI?
- 27:51What was the operation that was called? Things
- 27:55like that.
- 27:56Those are logs,
- 27:58and logs are strings.
- 28:00And if it's a string, it's always gonna
- 28:02be in a resource log if it's coming
- 28:04from, like, a native service perspective. So if
- 28:06you're just looking to rationalize, hey, do I
- 28:08need to turn on resource logs? If the
- 28:10thing you want to interrogate sounds and feels
- 28:13like and it is a string, yeah, you
- 28:15need resource logs to make it work. We
- 28:17talked about the activity log for Arm. That
- 28:19absolutely
- 28:20is a bunch of strings that you can
- 28:22go in query
- 28:23through Azure Monitor
- 28:25and light up scenarios around that without needing
- 28:27to enable resource logging on a given resource
- 28:30and standing up a log analytics thing or
- 28:33dumping your logs to a storage account and
- 28:34then worrying half, how do I download it?
- 28:36What's the scheme of the JSON file?
- 28:39All those kinds of things
- 28:41that come along with it along the way.
- 28:43Yes.
- 28:43Absolutely. Your audio wasn't going through the Discord.
- 28:46That was my fault. Yeah. Teams. It goes
- 28:48back to my Teams issue at the beginning.
- 28:50But, yeah, absolutely.
- 28:52Excellent points on all of that. Trying to
- 28:54think. Activity logs. We got platform metrics. We
- 28:57have resource logs. Which did you know? We
- 28:59have all that log data, Entra. Yes. And
- 29:02I'm gonna tie this back. The log data
- 29:03in Entra has been there. I noticed it
- 29:06with
- 29:07global secure access and some of the enhanced
- 29:09logging you can do now, because this is
- 29:11a requirement for enhanced logging, is you can
- 29:14actually configure Entra diagnostic settings. So these resource
- 29:18logs we just talked about, Entra has resource
- 29:20logs where you can go set up Entra
- 29:24to
- 29:25send its essentially diagnostic
- 29:28logs into log analytics as well. Mhmm. I
- 29:31don't know how long that's been there, but
- 29:32I haven't always had that set up because
- 29:34I haven't had a need to until I
- 29:36was playing with Global Secure Access. So for
- 29:38these things like monitor, entry ID, or like
- 29:41the activity log, entry ID logs, things like
- 29:43that, where they're there by default, but maybe
- 29:46they're not pumping their data out to to
- 29:48another source like Log Analytics. The reason that
- 29:51you start to pump the data out to
- 29:52Log Analytics, or maybe you go and you
- 29:54look at security specific offerings like Sentinel or
- 29:58things like that, that can ingest that data
- 30:00for you, or they can act as a
- 30:01target or a sync source for that data,
- 30:03is because you wanna retain it longer, or
- 30:05because you wanna start to get that crosscut
- 30:07visibility
- 30:08and insights into,
- 30:10do I need to correlate an event that
- 30:12happened in enter ID to an event that
- 30:14happened in the activity log to ultimately an
- 30:17interaction on the data plane in one of
- 30:19my resources.
- 30:20That might be something like, do we want
- 30:22the ability to track the malicious sign in?
- 30:24Say there was like a phishing hack, and
- 30:26somebody got in with some admin credentials to
- 30:29your enter ID tenant, And then from there,
- 30:32they created another user, they created a service
- 30:34principal, and they gave that some kind of
- 30:36elevated rates in your environment,
- 30:38and then they connected with that user. Great.
- 30:40The login attempt up here is in Entra,
- 30:43and then the other interaction now is down
- 30:45in the activity log, and then all the
- 30:47way down to what they do while they
- 30:49were there. Oh, I can see that they
- 30:51went into
- 30:53my storage account, and they changed the configuration
- 30:55of a container on my storage account, and
- 30:57they enabled it for public access, like public
- 31:00and on access. And then I tie that
- 31:02back to a metric, and I see that
- 31:03they egress some data or something like that
- 31:06along the way. You know, so it's, like,
- 31:09a little convoluted. I I I think, if
- 31:11you're not, like, deep and
- 31:13and, like, familiar with it and you haven't
- 31:15touched it all. But once you start to
- 31:17touch it all, it it starts to make
- 31:19sense because you're just decomposing it into the
- 31:21various layers.
- 31:23So really, you're walking through, like, the various
- 31:25API surfaces, right, like, Entra and Graph,
- 31:28down to
- 31:30Azure and Arm, and the control plane, ultimately
- 31:32maybe back to, like, data plane within an
- 31:34Azure service, if it happens to have a
- 31:36data plane.
- 31:37Not all services have data planes. Some of
- 31:39them only have control plane interactions.
- 31:42It's hit or miss depending on the service,
- 31:44so you do have to know what you're
- 31:45going for service by service. And someone in
- 31:47Discord was just asking to, like, being able
- 31:51to track who created, like, a function app
- 31:54and when they created it and stuff and
- 31:56where those would be. That would be more
- 31:57of that Azure activity log. That's gonna be
- 32:00logging more of those, and we had talked
- 32:03about a little bit more of those control
- 32:04plane creating those.
- 32:06But I do agree everything you just said
- 32:08about logging, and that's why I personally logging
- 32:11all of these to log analytics. We'll probably
- 32:13start talking about more security stuff in future
- 32:16episodes.
- 32:17To your point of
- 32:19it's these different APIs, it's these different control
- 32:22planes,
- 32:23but then somebody gets into NTRIC, gets into
- 32:25a storage account, finds a file out there
- 32:27that has other login data, and then they
- 32:29go log in to SharePoint and pull information
- 32:31out of SharePoint.
- 32:32It in my opinion, it's really important when
- 32:34you have this wide range of things. Like
- 32:37you said, bring all of those logs into
- 32:40one central spot. So when you are investigating
- 32:44an incident or doing some type of threat
- 32:46hunting or if you wanna set up some
- 32:48type of alerting,
- 32:49all these logs are together, and you're not
- 32:52jumping between the entry ID sign in log
- 32:54and then over to the SharePoint audit log
- 32:56and then over to the Azure activity log
- 33:00and then jumping back into your diagnostic logs
- 33:02for your storage account. It gets really hard
- 33:05to
- 33:06correlate everything when they're all in these disjointed
- 33:09places, and that's where I think log analytics
- 33:12and
- 33:13Sentinel, if we talk about Sentinel in the
- 33:14future,
- 33:16really give you a good security
- 33:18benefit when it comes to being able to
- 33:20pull all of this together. So it's an
- 33:22interesting conversation. I have it with customers
- 33:25quite a bit.
- 33:27There's
- 33:28the class of customers
- 33:30that is
- 33:32scrappy.
- 33:33I absolutely understand the motivations, where they're coming
- 33:35from, and they want these things to be
- 33:37free. In that world where they want them
- 33:40to be free, they kind of stick with
- 33:41the free offerings. So they do these things
- 33:43where they end up building bespoke tooling,
- 33:46say, to
- 33:48use the graph PowerShell commandlets
- 33:51to query
- 33:52the the sign in logs and things like
- 33:55that from Entra. And then they go and
- 33:58they wire up a CLI command to extract
- 34:01metrics from over here, and they build these.
- 34:05It's gonna come across as, like, derogatory. I
- 34:08don't mean it to be that way. You
- 34:09end up building these, like, Rube Goldberg machines
- 34:12that that are just these mismatches of things.
- 34:14And they accomplish their goals in many cases,
- 34:16like, they figured it out and they learn
- 34:18how to get there, and they think in
- 34:20their heads they did it on the cheap.
- 34:21But the reality is they had to invest
- 34:23all the time into building that bespoke tooling
- 34:25and doing all those things. And ultimately, you
- 34:28still need to spin up the compute. And
- 34:30and the compute is what costs you the
- 34:31money in these scenarios is, k, I need
- 34:33to wire something up that can be online
- 34:36and available with networking and disks so it
- 34:38can talk to that other thing and bring
- 34:40it down and ingest it. I caution folks
- 34:43about they get into their heads very quickly.
- 34:44They look at the pricing of Log Analytics
- 34:46or Sentinel or things like that, and they
- 34:48go, too much money. I can just go
- 34:49send my engineer off, and they can take
- 34:513 months and build it for me. Your
- 34:53engineer that just built that thing for 3
- 34:55months, depending on their salary, you might have
- 34:57been able to buy a year of Log
- 34:59Analytics at your ingestion rate. It all depends
- 35:02on your environment, like what rate you have
- 35:04and rate of change and churn and retention
- 35:06and those things that you want
- 35:09along the way. But,
- 35:10I I think if you take a step
- 35:11back and you peel it back, like some
- 35:13of these things that look like they cost
- 35:15a lot of money when it comes to
- 35:16security,
- 35:17they're like a net wash when it comes
- 35:19to
- 35:20resourcing
- 35:21and operations
- 35:22and things like that. You do have to
- 35:25keep those things in mind. Oh, yeah. Another
- 35:27interesting question in the chat. So if you
- 35:28use a third party SIEM instead of Sentinel,
- 35:31would you consider log analytics aggregation doubling up
- 35:34on a function? So
- 35:37you don't have to Maybe. Install. I I
- 35:39I think is the yeah.
- 35:41But I think this is another misnomer
- 35:44is, like, customers feel that they're
- 35:48locked in, and I I get it. To
- 35:49a certain degree, you are locked in. Right?
- 35:50If you're an Entree customer, absolutely. You're using
- 35:52Entree. You really don't have another choice there.
- 35:54That is the identity provider for Azure and
- 35:56M365.
- 35:58So you you can do things like have
- 36:00your IdP, Okta, or whatever, as the relying
- 36:02party, but the reality is you're using Entra
- 36:04at the end of the day. Some stuff
- 36:05you can't fight. Some of these things you
- 36:06can. Do you have to use Sentinel? No.
- 36:08Do you have to use Log Analytics?
- 36:10Absolutely not. You can do things like build
- 36:13out that integration
- 36:14for those diagnostic logs, which can include metrics
- 36:18in them. They're not just resource logs. You
- 36:19can pump out the metrics as well. You
- 36:21can build out those integrations around things like
- 36:23Event Grid and pump those out
- 36:26to any system that you want to. I
- 36:28have no expectation that every customer who comes
- 36:30to me who says, I'm a Splunk shop,
- 36:33is immediately going to go and turn the
- 36:36switch and learn Kusto and start doing Sentinel
- 36:39or log analytics tomorrow. That's not good for
- 36:41them. They're already a Splunk shop. Or you
- 36:43might have another solution that's out there. So
- 36:4690 ish percent of the time, and it's
- 36:48almost a 100% of the time with, like,
- 36:49the really large providers out there, they're going
- 36:52to have some integration
- 36:54that is going to be able to hook
- 36:55into
- 36:56Event Grid and the eventing system. That's a
- 36:59core service within Azure. Right? It's available. It's
- 37:02scalable.
- 37:03It it will do all those things for
- 37:05you, and it it would allow you to
- 37:07put that data out to where it needs
- 37:08to be. Your answer of maybe classic consulting
- 37:11101,
- 37:12but very true, there might be pieces of
- 37:15functionality
- 37:16that you could absolutely replicate and do yourself.
- 37:19It's just, to my point earlier, it's not
- 37:21worth your time, right? Like the engineering effort
- 37:23that you'll invest
- 37:24to duplicate some of these solutions when they've
- 37:26already been done and built for you, I
- 37:28I really do think you need to have
- 37:30that kind of rationalization
- 37:31moment
- 37:32of, do I want my engineer to take
- 37:35a day, a week, a month, a year,
- 37:37whatever it is to build this thing, or
- 37:40can I just buy it and be done
- 37:42with it? Let's be honest, like you're in
- 37:43a pay go service anyway, everything's not gonna
- 37:45be free for you. Be smart and pick
- 37:48and choose where you're spending your money. Yeah.
- 37:50And, right, it depends too is we're looking
- 37:52at this from a security lens. So I
- 37:54would say from a security lens,
- 37:57if you are sending everything to Log Analytics
- 37:59and everything to a third party SIEM,
- 38:02yeah, you're probably doubling up on a lot
- 38:04of stuff that you don't need to be.
- 38:06That being said,
- 38:07Log Analytics is also not just security. So
- 38:11I actually have
- 38:12some clients I've worked with as well that
- 38:14have 2 log analytics instances. 1 that they
- 38:16send everything to from a security standpoint or
- 38:20even certain,
- 38:21resource logs that they wanna look at from
- 38:23a security perspective, and then I have another
- 38:25one for, like, operations because, for instance,
- 38:29Azure app services.
- 38:31If you're hosting a website in there, part
- 38:33of the resource logs or those diagnostic logs
- 38:36are also used for app insights for doing
- 38:38things like tracking
- 38:40website visitors and response time on your website
- 38:43and things that
- 38:45operations developers,
- 38:47your SEO
- 38:48folks,
- 38:49if they care about response time, certain,
- 38:52depending on their role, may actually want some
- 38:55of these logs. So you may
- 38:57still want a certain aggregation in log analytics,
- 39:00not from a security perspective, but from a
- 39:04operations
- 39:05monitoring perspective or from a debugging
- 39:08perspective or from a we just wanna know
- 39:11if there's an error on one of our
- 39:12Windows servers, again, going back to some of
- 39:14the operations. So I think from a pure
- 39:16security perspective,
- 39:17yes. There can absolutely be use cases for
- 39:21sending data
- 39:22either to both or even, again, sending sometimes
- 39:26data to 2 different Log Analytics workspaces.
- 39:29We mentioned the diagnostic logs. When you configure
- 39:32those, you can send them multiple places. You
- 39:34can send them to 3 or 4 different
- 39:36locations. It's not, I'm only going to send
- 39:39these logs to 1 or the other. Yeah.
- 39:41It's like level of effort and things like
- 39:43that, and I was saying you've inquired the
- 39:44whole time. It's EventOps, mea culpa. So the
- 39:46other thing to keep in mind too is
- 39:48if you're a customer who's
- 39:51coming to Azure with an existing solution
- 39:54and you're, like, looking at this space and
- 39:56you're going, yeah, I know there's some native
- 39:58stuff, but maybe you are that customer who
- 40:00has a third party SIEM or something else.
- 40:04Maybe you use, like, a a different firewall.
- 40:06Like, you you're out there using FortiGates and
- 40:08you're, like, Azure Firewall is not my thing.
- 40:10Those options are often available to you as
- 40:12well. So it's not like you have to
- 40:14completely ditch the ecosystems
- 40:17that you're in today. There there there's a
- 40:19pretty wide swath of partners
- 40:23and ISV solutions
- 40:25that are available to you. So like in
- 40:28the monitoring space,
- 40:30Elastic's there. So you can do like Elastic
- 40:32Integrations, Elasticsearch, things like that. Datadog is another
- 40:36one that I run into a lot. Like,
- 40:38I end up working with a lot of
- 40:39our cloud native customers and things like that,
- 40:41like custom dev shops who are spinning things
- 40:43up. So I see a lot of that.
- 40:45You can do things like Event Hubs has
- 40:47native Kafka integrations.
- 40:49So if you're pumping data out and you
- 40:51wanna pump it through Kafka
- 40:53and then ingest that over in, say, Databricks
- 40:57or do some Spark analysis on top of
- 40:59it, that's just all available to you. It's
- 41:00all possible. It's there. The hooks are there.
- 41:03It's worth checking through the documentation
- 41:06and seeing,
- 41:07hey, is there something that's already here that's
- 41:09in my wheelhouse that I'm familiar with so
- 41:11that you don't need
- 41:13to reinvent the wheel. And I think that's
- 41:15a big consideration, right? Because if you're reinventing
- 41:18the wheel or you're at it net new
- 41:19and you've never done it before, it's not
- 41:21like an immediate security hole, but it's a
- 41:23gap for you. And the and the more
- 41:24gaps you have in that observability,
- 41:26the less comfort factor you have, and you
- 41:29start to go down the weird rabbit hole
- 41:30in your own head sometimes about it. Ultimately,
- 41:33like, the cloud is about plugging a a
- 41:35bunch of pieces together.
- 41:36I I always think about it as like
- 41:38Lego bricks and every if I gave you
- 41:40a bucket of Lego bricks and it had
- 41:42a 100 different colors in it and I
- 41:43just spilled it on the floor and I
- 41:44said build me a house, you're gonna build
- 41:46something that looks like a house.
- 41:48Your house might have 4 walls, and the
- 41:51one my overachieving son builds is gonna have
- 41:536, whatever it is. But it's still gonna
- 41:55be a house, and that's okay. You're not
- 41:58gonna necessarily land on the same solution that
- 41:59somebody else did. I see a lot of
- 42:00customers that are like, just tell me how
- 42:00the other customer did it. I'm like, somebody
- 42:01else did. I see a lot of customers
- 42:02who are like, just tell me how the
- 42:03other customer did it. I'm like, I I
- 42:05really don't want to.
- 42:06We're not gonna be bespoke,
- 42:08and it's not like everything's
- 42:10in, you know, a unique snowflake that fell
- 42:13on the ground,
- 42:14but
- 42:15at the same time,
- 42:17we're all gonna be a little bit different.
- 42:18And that's okay.
- 42:19That's the place we're gonna land. Yep. Agree.
- 42:22Scott, when we said at the beginning of
- 42:23this episode, we're gonna keep an eye on
- 42:25time and quit out 30 minutes. Yeah. And
- 42:27we didn't even make it past, like, metrics
- 42:29and resource logs. Who saw that coming? Not
- 42:31me. Maybe we need to rename this to,
- 42:33like, the Microsoft Cloud Security Podcasts.
- 42:37Something
- 42:38observability.
- 42:39Yeah. We can probably cut it here. I
- 42:41think that's a good kinda grounding
- 42:43and an overview
- 42:44of Log. Some of the observability
- 42:46pieces
- 42:47and some things to think about there. We
- 42:49should definitely come back and revisit some of
- 42:52the
- 42:53native tooling that's there. I I think it's
- 42:55worth talking about. Sentinel
- 42:57and and some of the other things there,
- 42:59Defender, which we've talked about Microsoft Defender in
- 43:01context of Lakeham 365. I don't know that
- 43:03we've ever talked about Defender for Cloud and
- 43:06some of the integrations that come on the
- 43:07Azure side. There's a bunch of third party
- 43:10stuff out there. So, yeah, we can just
- 43:11keep running with this one for a while.
- 43:13I think that might be the plan. We'll
- 43:15see how long it goes. Not how long
- 43:16we can drag it out. We don't wanna
- 43:18drag it out, but
- 43:19how long it takes us to cover it
- 43:21to our satisfaction?
- 43:222025.
- 43:23Here we come. Perfect. Alright. Scott, go enjoy
- 43:27your weekend.
- 43:28Sounds good. Rest. Relax. As always. Thanks, Ben.
- 43:31Thank you. Alright. Have a good one.
- 43:35If you enjoyed the podcast, go leave us
- 43:37a 5 star rating in iTunes. It helps
- 43:39to get the word out so more IT
- 43:41pros can learn about Office 365
- 43:43and Azure.
- 43:44If you have any questions you want us
- 43:46to address on the show or feedback about
- 43:48the show, feel free to reach out via
- 43:50our website,
- 43:51Twitter, or Facebook.
- 43:53Thanks again for listening, and have a great
- 43:55day.