SLAs for Azure Networking Services
Networking is a fundamental building block for all cloud services, and the SLAs reflect the necessity of robust networking performance.
- Azure Virtual Network: The SLA for a virtual network promises 99.99% availability to users, providing the backbone for most inter-service communication in Azure.
- Azure Load Balancer: A foundational service for ensuring the availability of applications, the Azure Load Balancer comes with a 99.99% SLA when configured for two or more virtual machines.
How Azure Calculates SLA Compliance
Implications of SLAs on Architecture and Design
The SLAs provided by Azure have significant implications for how systems are designed on the cloud. They drive the architecture decisions in terms of redundancy, resiliency, and recovery. For instance, understanding that the SLA for single VMs is lower than for multiple VMs in availability sets, architects design systems that leverage availability sets or zones to ensure higher uptime and meet their business requirements.
Moreover, architects need to consider SLAs at every layer of the application stack. While Azure may provide a high SLA for the database service, if the application itself is a single point of failure, the overall system availability will be lower than the database SLA suggests.
Service Credits as Part of SLAs
A common aspect of SLAs is the provision of service credits in case of SLA breaches. Service credits are a form of compensation offered to customers when the service levels fall below the promised thresholds. It’s important to note that these credits are not automatic; customers typically must file a claim for credits with detailed evidence of the SLA breach.
The amount of the service credit often depends on the amount of time the service was unavailable beyond the SLA. For example, a common tiered structure might offer a 10% credit if availability falls below 99.9% but above 99.0%, and a 25% credit if availability falls below 99.0%.