Resource-efficient software is intended to avoid unnecessary processing and storage when delivering the same output because reducing wasted computation contributes to performance and sustainability objectives; in this design the new outcome is to avoid tying service availability to one Availability Zone, which calls for Multi-AZ redundancy rather than Resource-efficient software or the earlier Avoid single points of failure.

Cost optimization pillar is intended to avoid unnecessary expenditure while achieving business outcomes because cost optimization focuses on understanding spend and delivering value at the right cost; in this design the new outcome is to avoid tying service availability to one Availability Zone, which calls for Multi-AZ redundancy rather than Cost optimization pillar or the earlier Avoid single points of failure.

Avoid single points of failure primarily supports the earlier need to remove a single component whose outage would take down the whole system because redundancy across failure boundaries helps keep a service available when one part fails, but the new task is to avoid tying service availability to one Availability Zone; that is why Multi-AZ redundancy is preferable here.

Avoid single points of failure is intended to remove a single component whose outage would take down the whole system because redundancy across failure boundaries helps keep a service available when one part fails; in this design the new outcome is to make system behavior visible through meaningful operational metrics and logs, which calls for Infrastructure observability rather than Avoid single points of failure or the earlier Vertical scaling.

Automate recovery is intended to restore normal application behavior without waiting for routine manual intervention because automated detection and recovery reduce the time needed to handle predictable failure conditions; in this design the new outcome is to make system behavior visible through meaningful operational metrics and logs, which calls for Infrastructure observability rather than Automate recovery or the earlier Vertical scaling.

Vertical scaling primarily supports the earlier need to increase the capacity of an existing compute resource because vertical scaling changes the size or power of one instance or server, but the new task is to make system behavior visible through meaningful operational metrics and logs; that is why Infrastructure observability is preferable here.

Multi-AZ redundancy addresses the separate need to avoid tying service availability to one Availability Zone because deploying suitable redundant resources in multiple AZs isolates many facility-level failures; it is not as direct as Horizontal scaling for the actual requirement to spread work across additional independent resources.

Small reversible changes addresses the separate need to reduce the blast radius and rollback effort of deployments because frequent controlled changes are easier to diagnose and reverse than large simultaneous releases; it is not as direct as Horizontal scaling for the actual requirement to spread work across additional independent resources.

Cost optimization pillar addresses the separate need to avoid unnecessary expenditure while achieving business outcomes because cost optimization focuses on understanding spend and delivering value at the right cost; it is not as direct as Horizontal scaling for the actual requirement to spread work across additional independent resources.

Security pillar is intended to protect information, systems and assets using risk-aware controls because the security pillar addresses identities, detection, infrastructure and data protection, and incident response; in this design the new outcome is to plan for individual components or zones becoming unavailable, which calls for Design for failure rather than Security pillar or the earlier Loose coupling.

Loose coupling primarily supports the earlier need to allow application components to fail or scale with less direct dependency on each other because queues and asynchronous communication can decouple component lifecycles, but the new task is to plan for individual components or zones becoming unavailable; that is why Design for failure is preferable here.

Operational excellence pillar is intended to improve how teams operate, monitor and continually refine workloads because operational excellence focuses on delivering business value through effective operations and continuous learning; in this design the new outcome is to plan for individual components or zones becoming unavailable, which calls for Design for failure rather than Operational excellence pillar or the earlier Loose coupling.

Horizontal scaling addresses a different objective, to spread work across additional independent resources, because horizontal scaling increases the number of workers or instances serving a workload; the checklist instead requires the distinct outcomes to restore normal application behavior without waiting for routine manual intervention; increase the capacity of an existing compute resource.

Capacity awareness addresses a different objective, to understand resource demand and quotas before workload demand exceeds limits, because good reliability planning accounts for capacity needs and service quotas; the checklist instead requires the distinct outcomes to restore normal application behavior without waiting for routine manual intervention; increase the capacity of an existing compute resource.

Operational excellence pillar addresses a different objective, to improve how teams operate, monitor and continually refine workloads, because operational excellence focuses on delivering business value through effective operations and continuous learning; the checklist instead requires the distinct outcomes to restore normal application behavior without waiting for routine manual intervention; increase the capacity of an existing compute resource.

Resource-efficient software is intended to avoid unnecessary processing and storage when delivering the same output because reducing wasted computation contributes to performance and sustainability objectives; in this design the new outcome is to avoid unnecessary expenditure while achieving business outcomes, which calls for Cost optimization pillar rather than Resource-efficient software or the earlier Sustainability pillar.

Sustainability pillar primarily supports the earlier need to minimize environmental impact associated with running workloads because sustainability emphasizes efficient consumption and reduction of energy and resource waste, but the new task is to avoid unnecessary expenditure while achieving business outcomes; that is why Cost optimization pillar is preferable here.

Multi-AZ redundancy is intended to avoid tying service availability to one Availability Zone because deploying suitable redundant resources in multiple AZs isolates many facility-level failures; in this design the new outcome is to avoid unnecessary expenditure while achieving business outcomes, which calls for Cost optimization pillar rather than Multi-AZ redundancy or the earlier Sustainability pillar.

That description belongs to Capacity awareness because good reliability planning accounts for capacity needs and service quotas; it is used to understand resource demand and quotas before workload demand exceeds limits, not to explain how Vertical scaling provides the function to increase the capacity of an existing compute resource.

That description belongs to Continuous performance tuning because performance-efficient architectures update technology choices and capacity based on evidence; it is used to revisit resource choices as application demand evolves, not to explain how Vertical scaling provides the function to increase the capacity of an existing compute resource.

That description belongs to Operational excellence pillar because operational excellence focuses on delivering business value through effective operations and continuous learning; it is used to improve how teams operate, monitor and continually refine workloads, not to explain how Vertical scaling provides the function to increase the capacity of an existing compute resource.

Increasing instance power does not resolve every account-level quota or dependency; that caveat is associated more closely with Capacity awareness for the need to understand resource demand and quotas before workload demand exceeds limits than with Experiment-driven performance choices for the requirement to benchmark workload alternatives before standardizing a configuration.

It differs from performance efficiency, which emphasizes appropriate resource use; that caveat is associated more closely with Reliability pillar for the need to help a workload recover from failures and continue functioning correctly than with Experiment-driven performance choices for the requirement to benchmark workload alternatives before standardizing a configuration.

It is not the same as purchasing the largest instance for peak theoretical demand; that caveat is associated more closely with Right-sized architecture for the need to match provisioned resources to measured workload requirements than with Experiment-driven performance choices for the requirement to benchmark workload alternatives before standardizing a configuration.

Operational excellence pillar is intended to improve how teams operate, monitor and continually refine workloads because operational excellence focuses on delivering business value through effective operations and continuous learning; in this design the new outcome is to make system behavior visible through meaningful operational metrics and logs, which calls for Infrastructure observability rather than Operational excellence pillar or the earlier Automate recovery.

Automate recovery primarily supports the earlier need to restore normal application behavior without waiting for routine manual intervention because automated detection and recovery reduce the time needed to handle predictable failure conditions, but the new task is to make system behavior visible through meaningful operational metrics and logs; that is why Infrastructure observability is preferable here.

Avoid single points of failure is intended to remove a single component whose outage would take down the whole system because redundancy across failure boundaries helps keep a service available when one part fails; in this design the new outcome is to make system behavior visible through meaningful operational metrics and logs, which calls for Infrastructure observability rather than Avoid single points of failure or the earlier Automate recovery.

AWS Secrets Manager primarily supports the earlier need to securely store supported credentials and coordinate secret rotation because secrets Manager reduces the need to embed passwords and tokens in application source code, but the new task is to identify changes in supported AWS resource settings over time; that is why AWS Config is preferable here.

Amazon Macie is intended to identify potentially sensitive data stored in Amazon S3 because macie discovers and classifies sensitive information in supported S3 data; in this design the new outcome is to identify changes in supported AWS resource settings over time, which calls for AWS Config rather than Amazon Macie or the earlier AWS Secrets Manager.

AWS Shield Standard is intended to provide baseline DDoS protections for supported AWS resources because shield Standard supplies automatic DDoS protection to supported resources; in this design the new outcome is to identify changes in supported AWS resource settings over time, which calls for AWS Config rather than AWS Shield Standard or the earlier AWS Secrets Manager.

AWS Certificate Manager TLS certificates is intended to provision and manage certificates for supported TLS endpoints because aCM manages public or imported certificates for use with integrated services; in this design the new outcome is to identify software vulnerabilities and unintended network exposure on supported workloads, which calls for Amazon Inspector vulnerability findings rather than AWS Certificate Manager TLS certificates or the earlier Amazon VPC Flow Logs traffic metadata.

AWS KMS managed keys is intended to control encryption keys for supported AWS integrated services because kMS supports managed cryptographic key lifecycle and access controls for integrated workloads; in this design the new outcome is to identify software vulnerabilities and unintended network exposure on supported workloads, which calls for Amazon Inspector vulnerability findings rather than AWS KMS managed keys or the earlier Amazon VPC Flow Logs traffic metadata.

Amazon VPC Flow Logs traffic metadata primarily supports the earlier need to record selected IP network-flow information for VPC resources because VPC Flow Logs capture metadata about accepted or rejected traffic at supported interfaces, but the new task is to identify software vulnerabilities and unintended network exposure on supported workloads; that is why Amazon Inspector vulnerability findings is preferable here.