Azure Advisor - Simply Explained

Hello everyone and welcome to another episode of Knowledge Nuggets today's topic, Azure Advisor. What exactly is it? Is it just another dashboard with a list of suggestions you ignore or is it something more useful? Most people think it's a passive recommendation tool, something that sits in the corner of the Azure portal and occasionally tells you things you already know. But that's not what it is. By the end of this episode, you'll understand what Azure Advisor actually does, how its five pillars work and how the Advisor score helps you track your progress. Grab your coffee, let's dive in. What exactly is Azure Advisor? Azure Advisor is a free, built-in cloud consultant that analyzes your entire Azure environment. It doesn't just sit there. It actively scans your resources and compares them against Microsoft's best practices. Think of it like a building inspector who walks through your setup with a checklist pointing out what needs fixing. Here's the thing, it covers five areas. Cost, security, reliability, performance and operational excellence, and it's completely free. No extra cost, no installation, nothing to configure. The recommendations are personalized to your environment, not generic advice from a blog post. They're specific to the resources you're actually running, the way you're running them, and the patterns it sees in your usage. It's integrated directly into the Azure portal, available from the moment you start using Azure, so you don't need to enable it, sign up or configure anything. It's just there. So how does it actually do its job? How Azure Advisor does its job? Think of Azure Advisor like a building inspector for your cloud environment. It walks through your resources, checks how you're using VMs, databases and storage, and then compares everything against the Azure well-architected framework. Microsoft's own set of best practices for building reliable, secure and efficient workloads. And it does all of this automatically behind the scenes. Now here's the thing, it doesn't just give you a report right away. It needs at least seven days of data before it can say anything meaningful. Machine learning helps identify patterns like resources that are underused or traffic spikes that look unusual. Each recommendation comes with an impact score, high, medium or low. High impact ones are the ones you want to tackle first because they give you the biggest return for your effort. There's also a quick fix feature. Instead of going through each resource one by one, you can apply multiple recommendations with just a few clicks. An advisor sends proactive alerts about new recommendations and upcoming service retirements, so you're never caught off guard when a service you depend on gets deprecated. So what are these five pillars? Let's take a look. The five pillars, a quick overview. Azure Advisor breaks its recommendations into five categories. Cost recommendations help you find savings, right sizing VMs using reserved instances, shutting down idle resources, security recommendations, close vulnerabilities, enabling multifactor authentication, restricting network access, encrypting data, reliability recommendations keep services running, backups, redundancy, availability zones, performance recommendations make apps faster, optimizing throughput, eliminating bottlenecks, database tuning, and operational excellence recommendations streamline processes governance monitoring, staying current with supported versions. Each pillar gets its own section in the advisor dashboard with tailored recommendations. And these pillars come directly from the Azure well architected framework, which is Microsoft's structured approach to building workloads that are cost effective, secure, reliable, performant and well managed. Let's break down each one starting with the most popular and cost cost optimization. The cost pillar is the one most people open first. Everyone wants to save money, right? So Advisor looks at your environment and finds idle or underutilized resources, things like virtual machines running 24/7 with single digit CPU usage, where you're paying for compute cycles you're not using. That's where the savings start. It suggests right sizing, meaning maybe you're paying for a large VM when a smaller one handles the load just fine. An advisor tells you exactly which SKU would work better. Advisor also recommends reserved instances, which can save you up to 72% compared to pay as you go pricing. If you know you'll need a certain VM size for a year or three, you lock in a lower rate. There's also Azure savings plans for compute. These are more flexible than reserved instances because they apply across different VM sizes and regions, but the discount is slightly lower. And as your hybrid benefit, let's you reuse on-premises, Windows Server and SQL Server licenses you've already paid for by bringing your own license to Azure and getting a reduced rate. You can adjust the look back period in advisor from 7 to 90 days, so the analysis reflects your actual usage patterns, not just a snapshot. And here's a real example, shutting down dev and test VMs after hours. Organizations doing this often reduce spending by 30% or more. It's a simple fix, just turn things off when nobody's using them. Saving money is great, but you also need to keep your environment secure. Security. Security recommendations in Azure Advisor come from Microsoft Defender for Cloud, and are based on the Microsoft Cloud Security benchmark, which aligns with industry standards like CIS and NIST. The most common high priority recommendation, enable multi-factor authentication for privileged accounts. It's the single most effective thing you can do to prevent account compromise. Next, restrict network access to storage accounts. Don't leave them open to the internet. Instead, configure firewall rules and service endpoints. Enable encryption for data address and in transit. Advisor will flag storage accounts, databases, and key vaults that aren't properly protected. Each recommendation has a score impact that raises your subscription secure score, which is a percentage that tells you how well you're following security best practices. Microsoft maintains over 100 compliance certifications, so when Advisor tells you to enable something, it's often because that configuration maps to a specific compliance requirement. And QuickFix works here too, letting you apply security remediations across multiple resources at once, especially valuable for smaller teams, without dedicated security staff. Once you're secure in saving money, the next concern is keeping things running. Reliability? The reliability pillar keeps your services running when things go wrong. Because things will go wrong, hardware fails, and networks go down, entire regions have outages. The real question is whether your workload survives it. And that's what Advisor helps you answer. The biggest recommendation, turn on availability zones for your critical resources. Availability zones are physically separate data centers, inside and as your region. Each one has its own power, cooling, and networking. So if one zone goes down, your workloads keep running in another. No manual intervention needed. It's built into the platform. Zone redundant storage is another big one. Think of it as having a copy of your data in a separate building. If one building burns down, the other copy is safe. Your data gets copied across multiple zones. If one zone fails, your data stays safe. You don't lose a thing. For databases, Advisor suggests georeplication. That's copying your database to a different Azure region, hundreds of miles away. That means if an entire region goes down, say a natural disaster takes out a whole Azure region, your database is still running in another region somewhere else. For app service, Advisor tells you to set a minimum instance count. At least two instances for production workloads. That way, if one instance fails or needs maintenance, the other one keeps handling requests. Same logic applies to auto scaling. You want your app to handle traffic spikes without crashing. Backups are a huge part of this pillar. Advisor checks whether you've set up proper backup strategies for your VMs, databases, and app services. If you haven't, it'll tell you. And it's not just about having backups. It's about testing them. A backup you've never restored isn't a backup. It's a hope. Advisor will remind you to test your restore process regularly. There's a newer feature here too. As of May 2025, Advisor now gives Zonal Resiliency recommendations for AKS clusters, VPN gateways, and Azure Firewall. So if you're running Kubernetes or have a VPN gateway, Advisor can tell you whether you need to spread those across zones for better uptime. It's a simple check that can save you from a major outage. Here's the real world example. Imagine one Azure zone in your region goes down. If you followed Advisor's reliability recommendations, your workloads keep running in another zone. Your users don't notice anything. That's the goal. Reliability isn't just about being up. It's about being resilient. Reliability keeps you online. Performance keeps you fast. Performance. The performance pillar makes your applications faster and more responsive for users. Nobody likes a slow app, slow apps, frustrated users, and hurt your business. And in the cloud, performance problems usually come from resources that are undersized or misconfigured. Advisor spots these issues before they become problems. The most common recommendation, right size VMs that are consistently overutilized. Advisor looks at your VM metrics over at least seven days. If CPU is consistently above 90% or I/O is above 95%, it flags that VM for an upgrade. You're hitting the limits of what that scale you can handle and your users are probably feeling it. Right sizing gives you more headroom without overpaying. There's database tuning too. Advisor pulls recommendations from SQL Database Advisor, which analyzes your query patterns and suggests index changes or query optimizations. Sometimes a single missing index is the difference between a query that takes two seconds and one that takes 200 milliseconds. That's a huge difference for user experience. Intelligent scaling is another area. Advisor suggests scaling strategies that match your actual traffic patterns. Instead of over-provisioning to handle peak load, which costs you money during quiet periods, you scale dynamically. Traffic goes up, resources scale up. Traffic goes down, resources scale down. That's the smart way to handle variable traffic. For traffic management, Advisor recommends as your front door and traffic manager to root users to the closest or fastest endpoint that reduces latency for users, far from your primary data center. And for caching, it suggests Azure Cash for Redis to take load of your databases. Frequently access the data stays in memory instead of hitting the database every time. There's also a recommendation to upgrade to modern SKUs that support accelerated networking. That feature gives your VMs lower latency and higher throughput by bypassing the host software for network traffic. It's a simple configuration change that makes a noticeable difference. An advisor will tell you if your current SKU supports it. The goal of the performance pillar is simple. Eliminate bottlenecks before your users notice them. By the time users complain about slow load times, you've already lost trust. Advisor helps you catch those problems early. That's the performance pillar in a nutshell. The last pillar is about keeping your environment clean and well managed. Operational excellence. The last pillar is operational excellence and honestly it's the one most people overlook. It's not as exciting as saving money or preventing a security breach, but it's the pillar that keeps everything else running smoothly. Think of it as the housekeeping and governance pillar and it's all about process and efficiency. How do you deploy resources? How do you monitor them? Are you using supported versions of the services you depend on? Operational excellence answers all of that. A big recommendation here is to streamline your deployment processes with Azure Policy. Instead of manually checking that every resource follows your standards, you define policies once and Azure enforces them automatically. Want to make sure nobody creates a storage account without encryption? Azure Policy handles that. Monitoring is another big area and advisor will tell you to enable traffic analytics so you can understand your network flows. Who's talking to what are their unusual patterns? Without this visibility, you're flying blind and won't know if something's wrong until a user tells you. Upgrading to modern API versions is a recurring theme because older APIs get deprecated and if you're still calling them, your automation scripts and tools will eventually break. Advisor flags these before they become a problem. Same thing with web application, firewall policies. If you haven't configured them properly for better protection, advisor will tell you. Service retirement alerts are a huge part of this pillar. Advisor monitors the services you're using and tells you when they're being deprecated. Classic storage accounts are retiring and cloud services classic is being retired. If you're running on something that's about to go away, advisor gives you a heads-up so you can plan your migration. There's also a really practical one, tracing enabled on production API management instances. That might sound harmless, but tracing logs can expose sensitive information like keys, access tokens, and passwords. Advisor catches that and tells you to turn it off in production. The point of operational excellence is simple. Keep your environment clean, manageable, and up-to-date. It's not glamorous, but it's the difference between a cloud environment that runs smoothly and one that's a constant fire drill. Now, how do you measure how well you're doing across all five pillars? The advisor score. The advisor score is your optimization dashboard, a single number from 0 to 100%, that shows how well your workloads follow Azure best practices across all five pillars. Instead of digging through dozens of recommendations and trying to figure out if you're making progress, you get one number broken into five category scores, one for each pillar. So you can see at a glance that your cost score is 92%, but your reliability score is only 65%, and that tells you exactly where to focus your attention. How is the score calculated? It uses the healthy resource ratio, the percentage of resources that follow the recommendations. If you have 100 VMs and 80 of them, follow the cost recommendations your cost score is 80%. Simple. The overall score is the average of all five category scores. So if you're scoring 80 in cost, 70 in security, 60 in reliability, 90 in performance, and 75 in operational excellence, your overall score is 75%, that's your baseline. For reliability, performance, and operational excellence, the scores are broken into subcategories with fixed weights. For example, zone resiliency carries a weight of 30 in the reliability category. Regional resiliency is 25, and data protection and recovery is 20. Each subcategory contributes proportionally to the category score, which means not all recommendations are equal. Some matter more than others, and the score reflects that. The score updates every 24 hours, so if you implement a recommendation today, it takes up to a day for your score to reflect that change. Don't expect instant gratification, but over time, you'll see your score trending upward as you work through the recommendations. It also tracks progress over time, so you can see daily, weekly, and monthly trends. You can look back and say, "Three months ago, my overall score was 62%, now it's 78%, here's the proof that we're improving." This is how you prove to your boss that you're making progress. Instead of saying, "I think we're doing better," you show them a chart. The advisor score gives you a concrete, measurable way to track optimization. If a recommendation doesn't apply to your situation, you can postpone or dismiss it, and those recommendations are excluded from the score calculation. So you're not penalized for ignoring something that doesn't make sense for your workload. So how do you actually start using all of this? Getting started with Azure Advisor. So you want to get started with Azure Advisor? Here's the thing. You don't need to install anything or configure anything. Just open the Azure portal, and advisor is right there in the navigation. You'll see personalized recommendations immediately no setup required. Start with the advisor score. That single number tells you where you stand today. Look at your overall score, then check which pillar is dragging you down. Focus on the lowest score first. That's where you'll get the biggest improvement with the least effort. Next, look at high impact recommendations. These are the ones that give you the most bang for your buck. Tackle those first. Don't get bogged down in low impact recommendations that barely move the needle. You want to save time and energy. Set a regular review cadence. Weekly or monthly. Put it on your calendar. Your environment changes constantly. New resources get added. Usage patterns shift. Services get deprecated. If you only check advisor once, you're missing the point. Think of it like a regular check-up for your cloud setup. Now here's where it gets powerful. Combine advisor with Azure Policy to enforce standards automatically. When advisor tells you something is wrong, create a policy that prevents it from happening again. That turns one time fixes into ongoing compliance. If you're more advanced, use the rest API and CLI tools to integrate recommendations into your DevOps workflows. You can pull recommendations programmatically and act on them as part of your deployment pipeline. That's the automation layer. Microsoft Learn has free training resources if you want to go deeper. But honestly, you don't need training to get started. Open advisor check your score, act on one high impact recommendation this week. That's it. Starts more. Pick one pillar. The rest will follow. So now you have the full picture. As your advisor is a free built-in cloud consultant that analyzes your entire environment across five pillars. Cost, security, reliability, performance, and operational excellence. It gives you personalized recommendations based on your actual usage, not generic advice. And the advisor score gives you a single number to track your progress over time. Your homework is simple. Open as your advisor today. Check your score and act on one high impact recommendation this week. That's it. Start small, but start. If this episode helped you understand Azure Advisor a little better, subscribe on your favorite podcast platform and share this with someone who's just starting their Azure journey. They'll thank you later. I'm Mirko Peters from M365. FM, thanks for listening.