Article Summary:
Our thoughts on the retirement of Azure Lab Services and the development of an Azure Labs alternative platform.
Introduction
For several years, Azure Labs has been a central component of technical training courses, assisting training providers wanting to add cloud-based practical activities to the delivery of their IT programmes.
Here at Kandu, the platform has been a key component of our BTEC Level 2 – IT Support and BTEC Level 3 – Networking & Cyber Security courses for a long time.
As a company, our goal has always been to enable learning outside of the traditional classroom and we have been successful in designing programmes which are remotely accessible to those who find scheduled learning difficult, which includes home educators and adult learners in full time employment.
Azure Labs is one of the key tools which has allowed us to do this and having the ability to spin-up a virtual machine and assign this to a student who can then access the resources as needed, from any location has been invaluable.
Like many other training providers, we were concerned by Microsoft’s announcement back in November of 2025 that the platform would be retired permanently in June of 2027.
The full announcement can be seen here: https://learn.microsoft.com/en-us/azure/lab-services/retirement-guide
What was great about Azure Labs?
Cost – Obviously, this is a huge factor. The cost of cloud computing can be terrifying and for a training provider the price of the resources which are needed to deliver a course will ultimately be passed on to the customer, whether that be a privately funded student or someone who has enrolled as part of a funded training scheme.
The Azure labs pricing model was very clear, each VM incurs a fixed price whilst it is running, when switched off there are no charges.
This paired with management features such as quotas, scheduling and automatic shutdown meant that it was possible to accurately predict the cost of using the labs on a per-student or per-group basis depending on the delivery type.
Freedom – Azure labs provided the platform for training providers like us to create. A blank slate which with a bit of imagination and technical know-how could be adapted to create real-world practical environments for all different types of IT training. We successfully developed lab activities for CompTIA certifications including the Network+ and Security+, several Microsoft Certifications and UK accredited qualifications from awarding bodies such as City & Guilds and BTEC.
There are many off-the-shelf lab or simulation solutions available, but using such products means we need to fit the course delivery around the supplied materials, rather than have the capacity to build such resources around the students.
Scalability/Elasticity – As is the nature of cloud-based resources, Azure Labs allowed a training provider to be up and working with minimal outlay and no significant investment in infrastructure.
Aside from some CPU core & storage limitations from Microsoft, it was very easy to cater for individual learners through to very large groups.
What was not-so-great about Azure Labs?
Performance – Publishing a template takes around 1 hour. Creating or re-imaging a Virtual Machine can take 30-40 minutes, starting a Windows based VM takes up to 5 minutes.
This was often a source of frustration for eager students.
Limited Troubleshooting Options – Azure Labs virtual machines are self-contained within the Azure Labs platform, which means that the VM’s generated by the template could not be accessed in the same way a VM can when using Azure Compute.
A common scenario would be for a learner to accidentally adjust the network settings of their host VM, cutting it off from the network leaving it inaccessible with no option of reverting the settings via the Cloud CLI as we would a regular compute based Virtual Machine. The only option in this situation was to re-image the system meaning the student would need to start their task from scratch and wait around 40 minutes until the re-image is completed.
Limited Storage – VM’s are created with a set maximum of 127 GB storage space available, this includes the Operating System. For a basic Windows installation this would suffice but if making use of nested virtualization for more complicated scenarios this would quickly become a barrier to what could be created. Running out of storage space on the platform has become common and often the only solution is to re-image (with another 40-minute wait).
Lack of Pre-built Templates – In our list of Azure Lab strengths, ‘Freedom’ was number 2 and being able to build anything you like (within the 127 GB limit) is a key feature, however an option to install some pre-built templates is a feature that could save some time.
Azure Labs Alternatives?
Azure Compute/DevTest Labs/Virtual Desktop
The obvious place to start is Azure itself. Virtual Machines can be created via Compute, DevTest Labs and Azure Virtual Desktop.
The first issue to consider here is management of the users. How does a tutor create the VM’s, then make these easily accessible to the students? How do we set quotas? How can we schedule access?
All of these things are possible but not ‘out-of-the-box’ as is the case with Azure Labs.
Then comes the scary part …. cost!
Outside of the Azure labs comfort zone, virtual machines attract many different charges which are often unexpected to those new to cloud computing.
First, we have the compute cost of the VM (price per hour to run). In addition to this there are the network bandwidth costs, the cost of having a public IP address needed for remote access and then a standing storage cost of $0.05 per GB per month. This may sound trivial at first, but can soon spiral out of control if not managed very carefully.
It is also extremely important to remember that charges still apply even if a Virtual Machine is ‘stopped’, you will still be charged for the resources which have been assigned to that VM.
We compared a scenario where a group of virtual machines on Azure Labs for 1 month cost approximately £120 to run.
The same scenario if hosted directly on the Azure Compute platform would cost an estimated £1600 due to the additional costs and standing charges!
Third Party Solutions:
In the retirement notice, Microsoft do make some suggestions with regards to partners who can help. We have investigated their offerings and found that although all provide excellent solutions, nothing met our specific needs for various reasons.
Most solutions require a per user subscription fee, starting at about £12 per user per month and rising upwards. This fee would be on top of paying your own Azure compute costs which outside of Azure Labs could be considerable.
Some partners include the cloud costs, but with this we lose the flexibility to create and as you might expect the per user costs is very high.
Some of the partner solutions also require a high guaranteed minimum spend per year which seems to be aimed at large organisations such as Universities, rather than small private training providers like us.
Build our own solution?
This was something we investigated and there are various approaches which could be taken such as building a solution on physical hardware which would have lower running costs, but a high initial outlay and limited elasticity in terms of the available resources.
Having seemingly exhausted all possibilities we were heading towards a hybrid solution, self-hosting certain lab activities and using off-the-shelf products for others.
The Perfect Solution?
One of the best things about delivering IT training since 2002 is the strong relationship we have developed with our former students, many of which have gone on to have stellar IT careers. One such former student is Nicholas Doropoulos.
Nicholas worked with us back at the start of his career completing BTEC Level 2, CompTIA A+ and Network+ certifications before going on to become a multi-award winning DevSecOps/Cloud Security Engineer, working for prestigious organisations such as AWS Shield and the European Parliament.
Nicholas now leads a team of talented DevOps engineers who have developed an Azure Labs Replacement platform.
For full disclosure, Kandu are a paying customer of Labiteers, with no financial interest in the project, nor have we received any incentives to write about the project here. As early adopters we have been fortunate to influence the design process based on our experience of using Azure Labs to design activities for us and other training providers.
We have also volunteered our services for product testing, motivated purely by our belief in the project and the need for a perfect replacement which will allow us to continue providing the best experience possible to our learners beyond June 2027.
This week sees an important milestone in our use of the platform as we have just replaced some of the Azure Labs based tasks in our Level 3 networking course with Labiteers based tasks instead. See our student guide below.
We will go into greater detail about the project in future posts, but as quick summary of some key features of Labiteers.
Cost Management – Labiteers must be connected to an Azure subscription. The product has been designed by Cloud Experts from the ground up with cost management being the primary consideration. Like Azure Labs, Labiteers incorporates an Auto-Shutdown feature to reduce running costs should a student forget to switch off their system. Further to this, an auto-delete feature is included, which when set can destroy a VM entirely if inactive for a set period, eliminating any standing charges.
Performance – The process of generating a VM takes around 3 minutes, a huge improvement on the 40-60 minutes we see with Azure Labs. The process of starting up a VM can be as little as 20 seconds for Linux or 1-2 mins at the most for Windows.
Complete Control – The Labiteers platform makes managing the VM’s easy for anyone, however, as the VM’s are not contained within a specific application as per Azure Labs, you can access advanced features of individual VM’s via the Azure portal. For example, if you need to open specific firewall ports for an advanced task, or troubleshoot a VM using the Cloud CLI this is all possible.
Choice of VM’s – Both Windows and Linux VM’s are supported, however when using Labiteers we are no longer bound by the 127 GB storage limited per VM which is imposed by Azure Labs.
We can create tiny, super-cheap VM’s
Or something much more powerful.
The resource management features, when used properly can also help navigate issues caused by core limitations.
Build your own & pre-built templates – You have the option to build your own VM template completely from scratch, or purchase affordable pre-built templates from the Labiteers Hub.
Customer Service – Outstanding direct communication with experts. With Azure labs you are basically on your own!
Conclusion:
We still have until June 2027 to solve the Azure Labs retirement problem.
The Labiteers project, started in April 2025 has made enormous progress in just 10 months and as early adopters we have been lucky enough to have some say in the design process and see the roadmap of features which will be added in the coming year. If you are a training provider worried about what comes next, then I think it is certainly worth getting in touch with the Labiteers team to discuss your specific requirements.