Designing high availability for Azure compute requires choosing the right level of physical and logical isolation. Availability Zones are physically separate datacenters within a single region, each containing independent power, cooling, and networking systems. In contrast, Availability Sets isolate virtual machines (VMs) logically within a single datacenter by spreading them across different hardware racks, known as fault domains and update domains. Designing workloads with Availability Zones protects applications against entire datacenter outages, whereas Availability Sets only guard against local hardware or power supply failures on a specific rack.
To manage large groups of identical VMs automatically, architects use Virtual Machine Scale Sets configured in flexible orchestration mode. This mode allows Azure to automatically distribute scale set instances and their underlying disks across multiple Availability Zones to ensure high availability. When setting up these scale sets, you must configure key settings to control how instances are balanced and protected:
When an application cannot use scale sets, you can still achieve high availability by distributing individual VMs across three Availability Zones. To ensure the application remains resilient, you must deploy at least two VMs per zone and direct incoming traffic to them using an Azure Load Balancer. The operating system and data disks for these VMs must be zone-redundant to prevent data loss. Using zone-redundant storage (ZRS) disks prevents data loss during a zone failure because Azure automatically replicates the disk data across multiple zones.
Maintaining application uptime during updates requires careful coordination between the network load balancer and compute instances. Before deploying updates, check the Azure Service Health Dashboard to confirm there are no ongoing platform issues or planned maintenance events. Next, configure the load balancer health probes to drain traffic away from the specific zone undergoing maintenance, ensuring no users experience dropped connections. Perform rolling upgrades one zone at a time to keep the overall application capacity stable during the process.
To handle unexpected traffic spikes, define autoscaling policies that automatically add or remove VM instances based on metric thresholds, such as average CPU usage exceeding eighty percent. These policies monitor performance in real-time to adjust capacity up or down across all configured zones. The load balancer and its health probes work together with these scaling rules to ensure that newly created VMs only receive traffic after they are fully running and have passed their initial health checks.
Venture into the world of Azure Infrastructure, where design meets functionality. Harness your skills and gain mastery over complex cloud structures to ace the AZ-305 Designing Microsoft Azure Infrastructure Solutions exam!
Prepare and test your skills
Prepare and test your skills
Availability Zones are physically separate datacenters within a single region, each with independent power, cooling, and networking. Availability Sets isolate virtual machines logically within a single datacenter by spreading them across different hardware racks known as fault domains and update domains. Zones protect against entire datacenter outages, while Sets only guard against local hardware or power supply failures on a specific rack.
In flexible orchestration mode, you can choose best-effort zone balance to prioritize quick VM creation even if zones become uneven, or strict zone balance to force an equal number of VMs in each zone. You can also configure fault domain spreading to use max spreading to spread VMs across as many physical racks as possible, or align VMs with the fault domains of their managed disks.
Zone-redundant storage (ZRS) is a storage option that automatically replicates disk data across multiple Availability Zones. Using ZRS prevents data loss during a zone failure because Azure maintains copies of the data in different physical locations within the region, ensuring applications remain resilient when an entire datacenter experiences an outage.
During maintenance, you configure the load balancer health probes to drain traffic away from the specific zone undergoing updates, ensuring users experience no dropped connections. You then perform rolling upgrades one zone at a time while maintaining overall application capacity. The health probes automatically remove unhealthy instances from the network rotation and ensure newly created VMs only receive traffic after they are fully running and have passed their initial health checks.
StudyCerts
StudyCerts