The session explains Azure virtual machine high availability and how to reduce single points of failure in production environments.
Azure options can help move workloads toward higher availability targets such as 99.95%.
Scaling concepts
Scaling up or down changes the resources allocated to an individual VM and can help when a single workload is constrained by compute resources.
Scaling out or in adds or removes VMs to spread load across more instances when demand comes from many users or threads.
The right scaling approach depends on the actual resource constraint; vertical and horizontal scaling do not solve every problem.
Azure maintenance and downtime
Microsoft manages Azure infrastructure at massive scale, including hardware failures, patching, driver updates, and planned maintenance.
Azure uses techniques such as draining workloads and deploying freshly patched operating systems to reduce maintenance impact.
Update domains help Microsoft avoid updating multiple related nodes at the same time, minimizing impact on customer VMs.
Availability sets
An availability set groups two or more VMs, typically behind a load balancer, to improve availability within a region.
Availability sets use fault domains and update domains to spread VMs across infrastructure and reduce single points of failure.
A VM must be placed into an availability set when it is created; an existing VM cannot be added later.
Planning capacity must account for normal load and maintenance scenarios, because patching or rebooting one VM shifts traffic to the remaining VMs.
Availability sets can use up to three fault domains and up to 20 update domains.
Availability zones
Availability zones provide redundancy across separate data centers inside the same Azure region.
Zones are independent in areas such as power and cooling, helping protect applications from data center failures.
Users can select zones manually or allow Azure to distribute VMs evenly across zones.
Scale sets
Virtual machine scale sets use a master image and a load balancer backend pool to manage many VM instances as a group.
Scale sets can define minimum and maximum instance counts and can scale manually or automatically based on rules.
Scale sets can automatically add instances when load crosses thresholds, such as adding two VMs when usage reaches 70%.
Scale sets simplify updates by creating a new master image version, draining old instances, and replacing them with new version instances.
Portal demonstration
The demo shows how to create availability sets, availability zones, and scale sets in the Azure portal.
An availability set can be pre-built or created while creating the first VM that will use it.
When creating a VM, availability options include none, availability zone, virtual machine scale set, or availability set.
For availability zones, manual selection requires the administrator to balance placement, while Azure-selected zones distribute VMs evenly.
Scale set creation includes manual scaling and autoscaling options, with detailed scaling rules covered in another module.
Infrastructure management and related services
Azure Arc provides centralized cloud-based management for Azure resources, on-premises devices, and other cloud environments.
High availability in Azure also involves other services and patterns, including storage replication, region pairs, Cosmos DB, Traffic Manager, Front Door, containers, Kubernetes, and App Service plans.
The session focuses specifically on virtual machine availability within the broader set of Azure high availability capabilities.
Actiepunten
Watch the related recording on virtual machine self-healing for more detail on unexpected downtime handling.
Watch the App Service Plans module to learn scaling rules and how to manage them.
Continue learning by watching other videos in the course or exploring Microsoft Learn.
