A centralized observability store isn't recommended to avoid a single point of failure. Monitoring data for global resources and regional resources are stored independently. Have federated workspaces for observability data. Refer to Well-Architected mission-critical workloads: Deployment and testing. Have continuous validation by integrating automated testing as part of DevOps processes, including synchronized load and chaos testing, to fully validate the health of both the application code and underlying infrastructure. This eliminates risks associated with deployment and process variations across environments. Build and release pipelines must be fully automated to deploy stamps as a single operational unit, using blue/green deployments with continuous validation applied.Īpply environment consistency across all considered environments, with the same deployment pipeline code across production and pre-production environments. Implement zero downtime blue/green deployment pipelines. Refer to Well-Architected mission-critical workloads: Scale-unit architecture.Īpply the principle of Infrastructure as code (IaC) using technologies, such as Terraform, to provide version control and a standardized operational approach for infrastructure components. Each stamp also applies multiple nested scale units, such as the Frontend APIs and Background processors which can scale in and out independently. Refer to Well-Architected mission-critical workloads: Global distribution.ĭeploy a regional stamp as a scale unit where a logical set of resources can be independently provisioned to keep up with the changes in demand. Utilize Availability Zones (AZs) for all considered services to maximize availability within a single Azure region, distributing components across physically separate data centers inside a region.Ĭhoose resources that support global distribution.
The application is distributed across two or more Azure regions that handle active user traffic. This architecture applies a set of overarching design strategies intended to address these factors and ensure the target reliability tier is achieved.ĭeploy to multiple regions in an active-active model. Many factors can affect the reliability of an application, such as the ability to recover from failure, regional availability, deployment efficacy, and security. Refer to Well-Architected mission-critical workloads: Design for business requirements. These individual numbers should be aggregated to determine a composite SLA which should align with workload targets. To define a realistic SLO, it's important to understand the SLA of all Azure components within the architecture.
0 kommentar(er)
