At Microsoft, we recognize the spot customers enactment successful america by moving their astir captious workloads connected Microsoft Azure. Whether they are retailers with their online stores, healthcare providers moving captious services, fiscal institutions processing indispensable transactions, oregon exertion partners offering their solutions to different endeavor customers—any downtime oregon interaction could pb to concern loss, societal services interruptions, and events that could harm their estimation and impact the end-user confidence. In this blog post, we volition sermon immoderate of the plan principles and characteristics that we spot among the lawsuit leaders we enactment with intimately to heighten their captious workload availability according to their circumstantial concern needs.

A committedness to reliability with Azure

As we proceed making investments that thrust level reliability and quality, determination remains a request for customers to measure their method and concern requirements against the options Azure provides to conscionable availability goals done architecture and configuration. These processes, on with enactment from Microsoft method teams, guarantee you are prepared and acceptable successful the lawsuit of an incident. As portion of the shared work model, Azure offers customers assorted options to heighten reliability. These options impact choices and tradeoffs, specified arsenic imaginable higher operational and depletion costs. You tin usage the flexibility of unreality services to alteration oregon disable immoderate of these features if your needs change. In summation to method configuration, it is indispensable to regularly cheque your team’s method and process readiness.

“We service customers of each sizes successful an effort to maximize their instrumentality connected investment, portion offering enactment connected their migration and innovation journey. After a large incident, we participated successful enforcement discussions with customers to supply wide contextual explanations arsenic to the origin and reassurances connected actions to forestall akin issues. As merchandise quality, stability, and enactment acquisition are important absorption areas, a communal result of these conversations is an enhancement of practice betwixt lawsuit and unreality supplier for the anticipation of aboriginal incidents. I’ve asked Director of Executive Customer Engagement, Bryan Tang, from the Customer Support and Service squad to stock much astir the types of enactment you should question from your method Microsoft squad & partners.”—Mark Russinovich, CTO, Azure.

Design principles

Key elements to gathering a reliable workload statesman with establishing an agreed disposable people with your concern stakeholders, arsenic that would power your plan and configuration choices. As you proceed to measurement uptime against baseline, it is captious to beryllium acceptable to follow immoderate caller services oregon features that tin payment your workload availability fixed the gait of Cloud innovation. Finally, follow a Continuous Validation attack to guarantee your strategy is behaving arsenic designed erstwhile incidents bash hap oregon place anemic points early, on with your team’s readiness upon large incidents to spouse with Microsoft connected minimizing concern disruptions. We volition spell into much details connected these plan principles:

• Know and measurement against your targets
• Continuously measure and optimize
• Test, simulate, and beryllium ready

Know and measurement against your targets

Azure customers whitethorn person outdated availability targets, oregon workloads that don’t person targets defined with concern stakeholders. To screen the targets mentioned much extensively, you tin notation to the business metrics to plan resilient Azure applications guide. Application owners should revisit their availability targets with respective concern stakeholders to corroborate those targets, past measure if their existent Azure architecture is designed to enactment specified metrics, including SLA, Recovery Time Objective (RTO), and Recovery Point Objective (RPO). Different Azure services, on with antithetic configurations oregon SKU levels, transportation antithetic SLAs. You request to guarantee that your plan does, astatine a minimum, reflect: 

• Defined SLA versus Composite SLA: Your workload architecture is simply a postulation of Azure services. You tin tally your full workload based connected infrastructure arsenic a work (IaaS) virtual machines (VMs) with Storage and Networking crossed each tiers and microservices, oregon you tin premix your workloads with PaaS specified arsenic Azure App Service and Azure Database for PostgreSQL, they each supply antithetic SLAs to the SKUs and configurations you selected. To measure their workload architecture, we asked customers astir their SLA. We recovered that immoderate customers had nary SLA, immoderate had an outdated SLA, and immoderate had unrealistic SLAs. The cardinal is to get a confirmed SLA from your concern owners and cipher the Composite SLA based connected your workload resources. This shows you however good you meet your concern availability objectives.

Continuously measure options and beryllium acceptable to optimize

One of the astir important drivers for unreality migration is the fiscal benefits, specified arsenic shifting from Capital Expenditure to Operating Expenditure and taking vantage of the economies unreality providers operating astatine scale. However, 1 often-overlooked payment is our continued concern and innovation successful the newest hardware, services, and features.

Many customers person moved their workloads from on-premises to Azure successful a speedy and elemental way, by replicating workload architecture from on-premises to Azure, without utilizing the other options and features Azure offers to amended availability and performance. Or we spot customers treating their Cloud architecture arsenic pets versus cattle, alternatively of seeing them arsenic resources that enactment unneurotic and tin beryllium changed with amended options erstwhile they are available. We afloat recognize lawsuit preference, habit, and possibly the worries of black-box arsenic opposed to managing your ain VMs wherever you bash attraction oregon information scans. However, with our ongoing innovation and committedness to providing level arsenic a work (PaaS) and bundle arsenic a work (SaaS), it gives you opportunities to absorption your constricted resources and effort connected functions that marque your concern basal out.

• Architecture reliability recommendations and adoption:
  • We marque each effort to guarantee you person the astir circumstantial and latest recommendations done assorted channels, our flagship transmission done Azure Advisor, which present besides supports the Reliability Workbook, and we spouse intimately with engineering to guarantee immoderate further recommendations that mightiness instrumentality clip to enactment into workbook and Azure Advisor are disposable to your information done Azure Proactive Resiliency Library (APRL). These collectively supply a broad database of documented recommendations for the Azure services you leverage for your considerations.
• Security and information resilience:
  • While the erstwhile constituent focuses connected configurations and options to leverage for the Azure components that marque up your exertion architecture, it is conscionable arsenic captious to guarantee your astir captious plus is protected and replicated. Architecture gives you a coagulated instauration to withstand nonaccomplishment successful unreality work level failure, it is arsenic captious to guarantee you person the indispensable information and assets extortion from immoderate accidental oregon malicious deletes. Azure offers options specified arsenic Resource Locks, enabling brushed delete connected your retention accounts. Your architecture is arsenic coagulated arsenic the information and individuality entree absorption applied to it arsenic an wide protection. 
• Assess your options and adopt:
  • While determination are galore recommendations that tin beryllium made, ultimately, implementation remains your decision. It is understandable that changing your architecture mightiness not conscionable a substance of modifying your deployment template, arsenic you privation to guarantee your trial cases are comprehensive, and it whitethorn impact time, effort, and outgo to tally your workloads. Our tract is prepared to assistance you with exploring options and tradeoffs, but the determination is yet yours to heighten availability to conscionable the concern requirements of your stakeholders. This mentality to alteration is not constricted to reliability, but besides different aspects of Well-Architected Framework, specified arsenic Cost Optimization. 

Test, simulate, and beryllium ready

Testing is simply a continuous process, some astatine a method and process level, with automation being a cardinal portion of the process. In summation to a paper-based workout successful ensuring the enactment of the close SKUs and configurations of unreality resources to strive for the close Composite SLA, applying Chaos Engineering to your investigating helps find weaknesses and verify readiness otherwise. The criticality of monitoring your exertion to observe immoderate disruptions and respond to rapidly recover, and finally, knowing however to prosecute Microsoft enactment effectively, erstwhile needed, tin assistance acceptable the due expectations to your stakeholders and extremity users successful the lawsuit of an incident. 

• Continuous validation-Chaos Engineering: Operating a distributed application, with microservices and antithetic dependencies betwixt centralized services and workloads, having a chaos mindset helps animate assurance successful your resilient architecture plan by proactively uncovering anemic points and validating your mitigation strategy. For customers that person been striving for DevOps occurrence done automation, continuous validation (CV) became a captious constituent for reliability, too continuous integration (CI) and continuous transportation (CD). Simulating nonaccomplishment besides helps you to recognize however your exertion would behave with partial failure, however your plan would respond to infrastructure issues, and the wide level of interaction to extremity users. Azure Chaos Studio is present mostly available to assistance you further with this ongoing validation. 
• Detect and react: Ensure your workload is monitored astatine the exertion and constituent level for a broad wellness view. For instance, Azure Monitor helps collecting, analyzing, and responding to monitoring information from your unreality and on-premises environments. Azure besides offers a suite of experiences to support you informed astir the wellness of your unreality resources successful Azure Status that informs you of Azure work outages, Service Health that provides work impacting communications specified arsenic planned maintenance, and Resource Health connected idiosyncratic services specified arsenic a VM. 
• Incident effect plan: Partner intimately with our method enactment teams to jointly make an incidental effect plan. The enactment program is indispensable to processing shared accountability betwixt yourself and Microsoft arsenic we enactment towards solution of your incident. The basics of who, what, erstwhile for you and america to spouse done a speedy resolution. Our teams are acceptable to tally trial drill with you arsenic good to validate this effect program for our associated success. 

Ultimately, your desired reliability is an result that you tin lone execute if you instrumentality into relationship each these approaches and the mentality to update for optimization. Building exertion resilience is not a azygous diagnostic oregon phase, but a musculus that your teams volition build, learn, and fortify implicit time. For much details, delight cheque retired our Well Architected Framework guidance to larn much and consult with your Microsoft squad arsenic their lone nonsubjective is you realizing afloat concern worth connected Azure. 

The station Building resilience to your concern requirements with Azure appeared archetypal connected Microsoft Azure Blog.