Amazon AWS Certified Solutions Architect - Professional SAP-C02 Bundle

Although Move heavy historical analytics into Redshift or a suitable columnar analytics design can be valid when An analytical SQL report scans large rows in a relational OLTP database each hour, it targets a different technical condition. Columnar analytics engines optimize scans and aggregations while preserving OLTP capacity.

Rebalance shard routing and choose a healthier index-sharding and ingestion design would address a different requirement: An OpenSearch cluster has one overloaded shard while the rest are idle. Skewed shards cause hotspot latency regardless of aggregate cluster size.

This alternative, Move to a supported gp3 or io2 configuration with deliberately provisioned IOPS and throughput, is appropriate when An EC2-hosted database is limited by baseline EBS IOPS under peak transactional writes. Tunable SSD volume performance addresses block I/O limits independent of capacity.

For the separate situation An RDS Multi-AZ standby has been promoted but application clients keep connecting to a cached old database IP, Reconnect clients by resolving the stable RDS endpoint DNS after failover could be useful. RDS managed failover updates endpoint DNS while clients pinned to old IPs cannot reach the new writer.

Use highly available Route 53 Resolver endpoints and DNS-server redundancy solves the related scenario An internal app relies on a single custom DNS server on one EC2 instance; resolver endpoint and backend redundancy prevent one host failure from breaking name resolution.

Although Put stateless instances in a multi-AZ Auto Scaling group behind a load balancer can be valid when An application relies on one manually configured EC2 instance that cannot recover from host failure, it targets a different technical condition. Desired-capacity replacement makes an individual EC2 host disposable.

For the separate situation An internal app relies on a single custom DNS server on one EC2 instance, Use highly available Route 53 Resolver endpoints and DNS-server redundancy could be useful. Resolver endpoint and backend redundancy prevent one host failure from breaking name resolution.

Add an independent NAT gateway per used AZ with AZ-local routing solves the related scenario A web fleet spans three AZs but all egress traffic uses one NAT gateway in AZ A; zonal egress isolation avoids one NAT failure interrupting the entire web fleet.

Although Use a managed Multi-AZ ElastiCache replication group with automatic failover can be valid when Two microservices share a single self-hosted Redis node without persistence or replica failover, it targets a different technical condition. Managed replicas reduce single-node cache outage impact.

Put stateless instances in a multi-AZ Auto Scaling group behind a load balancer solves the related scenario An application relies on one manually configured EC2 instance that cannot recover from host failure; desired-capacity replacement makes an individual EC2 host disposable.

Use a managed Multi-AZ ElastiCache replication group with automatic failover would address a different requirement: Two microservices share a single self-hosted Redis node without persistence or replica failover. Managed replicas reduce single-node cache outage impact.

This alternative, Use highly available Route 53 Resolver endpoints and DNS-server redundancy, is appropriate when An internal app relies on a single custom DNS server on one EC2 instance. Resolver endpoint and backend redundancy prevent one host failure from breaking name resolution.

Although Put stateless instances in a multi-AZ Auto Scaling group behind a load balancer can be valid when An application relies on one manually configured EC2 instance that cannot recover from host failure, it targets a different technical condition. Desired-capacity replacement makes an individual EC2 host disposable.

Reconnect clients by resolving the stable RDS endpoint DNS after failover would address a different requirement: An RDS Multi-AZ standby has been promoted but application clients keep connecting to a cached old database IP. RDS managed failover updates endpoint DNS while clients pinned to old IPs cannot reach the new writer.

For the separate situation A web fleet spans three AZs but all egress traffic uses one NAT gateway in AZ A, Add an independent NAT gateway per used AZ with AZ-local routing could be useful. Zonal egress isolation avoids one NAT failure interrupting the entire web fleet.

Use a managed Multi-AZ ElastiCache replication group with automatic failover would address a different requirement: Two microservices share a single self-hosted Redis node without persistence or replica failover. Managed replicas reduce single-node cache outage impact.

This alternative, Add an independent NAT gateway per used AZ with AZ-local routing, is appropriate when A web fleet spans three AZs but all egress traffic uses one NAT gateway in AZ A. Zonal egress isolation avoids one NAT failure interrupting the entire web fleet.

For the separate situation An application relies on one manually configured EC2 instance that cannot recover from host failure, Put stateless instances in a multi-AZ Auto Scaling group behind a load balancer could be useful. Desired-capacity replacement makes an individual EC2 host disposable.

For the separate situation An Auto Scaling group repeatedly fails to obtain the same EC2 instance type during a regional capacity shortage, Use flexible mixed-instance policies across several AZs with approved capacity choices could be useful. Diversified instance selection reduces dependence on one capacity pool.

Expand subnet CIDR capacity and plan ENI/IP requirements for task scaling solves the related scenario A production ECS service cannot launch enough tasks because the VPC subnets run out of IPs; container network interfaces consume address capacity during elastic scale-out.

Introduce RDS Proxy or bounded client pooling with safe concurrency limits would address a different requirement: An RDS connection storm overwhelms the writer after many Lambda functions scale out. Database connection pools protect the backend from bursty client creation.

Model and request appropriate Lambda concurrency capacity while protecting downstream limits solves the related scenario A high-traffic service expects thousands of Lambda invocations but unexpectedly hits regional concurrency throttling; concurrency quotas can cap execution independently of workload demand.

Introduce RDS Proxy or bounded client pooling with safe concurrency limits would address a different requirement: An RDS connection storm overwhelms the writer after many Lambda functions scale out. Database connection pools protect the backend from bursty client creation.

This alternative, Tune predictive or target-tracking scaling metrics using message age and arrival patterns, is appropriate when A queue depth scales consumers only after processing lag exceeds a business SLA. Queue-age signals make capacity decisions sensitive to processing delay.

Although Use flexible mixed-instance policies across several AZs with approved capacity choices can be valid when An Auto Scaling group repeatedly fails to obtain the same EC2 instance type during a regional capacity shortage, it targets a different technical condition. Diversified instance selection reduces dependence on one capacity pool.

Tune predictive or target-tracking scaling metrics using message age and arrival patterns would address a different requirement: A queue depth scales consumers only after processing lag exceeds a business SLA. Queue-age signals make capacity decisions sensitive to processing delay.

This alternative, Model and request appropriate Lambda concurrency capacity while protecting downstream limits, is appropriate when A high-traffic service expects thousands of Lambda invocations but unexpectedly hits regional concurrency throttling. Concurrency quotas can cap execution independently of workload demand.

Tune predictive or target-tracking scaling metrics using message age and arrival patterns would address a different requirement: A queue depth scales consumers only after processing lag exceeds a business SLA. Queue-age signals make capacity decisions sensitive to processing delay.

This alternative, Expand subnet CIDR capacity and plan ENI/IP requirements for task scaling, is appropriate when A production ECS service cannot launch enough tasks because the VPC subnets run out of IPs. Container network interfaces consume address capacity during elastic scale-out.

For the separate situation An RDS connection storm overwhelms the writer after many Lambda functions scale out, Introduce RDS Proxy or bounded client pooling with safe concurrency limits could be useful. Database connection pools protect the backend from bursty client creation.

For the separate situation A high-traffic service expects thousands of Lambda invocations but unexpectedly hits regional concurrency throttling, Model and request appropriate Lambda concurrency capacity while protecting downstream limits could be useful. Concurrency quotas can cap execution independently of workload demand.

Expand subnet CIDR capacity and plan ENI/IP requirements for task scaling solves the related scenario A production ECS service cannot launch enough tasks because the VPC subnets run out of IPs; container network interfaces consume address capacity during elastic scale-out.

Although Introduce RDS Proxy or bounded client pooling with safe concurrency limits can be valid when An RDS connection storm overwhelms the writer after many Lambda functions scale out, it targets a different technical condition. Database connection pools protect the backend from bursty client creation.