Amazon AWS Certified Solutions Architect - Professional SAP-C02 Bundle

This alternative, Inspect CloudWatch Database Insights wait events and I/O saturation, is appropriate when A relational database's average CPU is low but client requests stall waiting on storage reads. Database wait analysis identifies I/O bottlenecks that CPU metrics alone can miss.

For the separate situation An RDS database's transactional reads grow rapidly while writes remain steady, Add appropriate read replicas and route read-only queries away from the writer could be useful. Read scaling isolates analytics and reads from transactional write load.

Add RDS Proxy and control client connection concurrency solves the related scenario A database experiences rising connection counts during Lambda traffic spikes; rDS Proxy pools application connections to reduce database connection pressure.

For the separate situation An application is throttled by EBS throughput despite spare vCPU and memory, Measure EBS bandwidth limits and select suitable instance and volume throughput could be useful. Both EC2 instance and EBS volume throughput caps constrain sustained I/O.

Although Co-locate latency-sensitive components appropriately and reduce synchronous round trips can be valid when A cross-zone chatty microservice introduces significant latency between dependent calls, it targets a different technical condition. Network hop and serialized RPC latency can outweigh individual function execution.

Scale on request or active connection metrics with prewarmed minimum tasks would address a different requirement: An ECS service scales slowly because capacity follows CPU but tasks queue on open connections. More representative demand signals reduce delayed scaling for connection-bound workloads.

Benchmark a memory-optimized EC2 family with sufficient RAM solves the related scenario An analytics workload with large in-memory joins repeatedly swaps to disk; eliminating swap pressure can improve join throughput more than clock-speed changes.

Scale on request or active connection metrics with prewarmed minimum tasks would address a different requirement: An ECS service scales slowly because capacity follows CPU but tasks queue on open connections. More representative demand signals reduce delayed scaling for connection-bound workloads.

This alternative, Profile CPU steal, I/O waits, network and thread contention across p99 traces, is appropriate when An EC2 API's p99 latency rises under load although average CPU remains under 30%. Tail latency may come from non-CPU resources or request concurrency bottlenecks.

Although Scale on request or active connection metrics with prewarmed minimum tasks can be valid when An ECS service scales slowly because capacity follows CPU but tasks queue on open connections, it targets a different technical condition. More representative demand signals reduce delayed scaling for connection-bound workloads.

Measure EBS bandwidth limits and select suitable instance and volume throughput would address a different requirement: An application is throttled by EBS throughput despite spare vCPU and memory. Both EC2 instance and EBS volume throughput caps constrain sustained I/O.

For the separate situation An analytics workload with large in-memory joins repeatedly swaps to disk, Benchmark a memory-optimized EC2 family with sufficient RAM could be useful. Eliminating swap pressure can improve join throughput more than clock-speed changes.

Profile CPU steal, I/O waits, network and thread contention across p99 traces would address a different requirement: An EC2 API's p99 latency rises under load although average CPU remains under 30%. Tail latency may come from non-CPU resources or request concurrency bottlenecks.

For the separate situation An application is throttled by EBS throughput despite spare vCPU and memory, Measure EBS bandwidth limits and select suitable instance and volume throughput could be useful. Both EC2 instance and EBS volume throughput caps constrain sustained I/O.

Co-locate latency-sensitive components appropriately and reduce synchronous round trips solves the related scenario A cross-zone chatty microservice introduces significant latency between dependent calls; network hop and serialized RPC latency can outweigh individual function execution.

This alternative, Measure EBS bandwidth limits and select suitable instance and volume throughput, is appropriate when An application is throttled by EBS throughput despite spare vCPU and memory. Both EC2 instance and EBS volume throughput caps constrain sustained I/O.

For the separate situation An analytics workload with large in-memory joins repeatedly swaps to disk, Benchmark a memory-optimized EC2 family with sufficient RAM could be useful. Eliminating swap pressure can improve join throughput more than clock-speed changes.

Although Profile CPU steal, I/O waits, network and thread contention across p99 traces can be valid when An EC2 API's p99 latency rises under load although average CPU remains under 30%, it targets a different technical condition. Tail latency may come from non-CPU resources or request concurrency bottlenecks.

Inspect cache-key headers, cookies and TTL policies to improve cacheability solves the related scenario A global media site retrieves the same files from origin for most requests despite a CloudFront distribution; excessive cache-key variation fragments edge cache hit rates.

Right-size the cache working set and inspect evictions and memory pressure would address a different requirement: A database cache contains all products but frequently accessed records still miss because cache memory is undersized. Capacity pressure can evict hot keys before they are reused.

This alternative, Add explicit invalidation or short TTL for critical mutable inventory fields, is appropriate when A product API returns stale inventory levels after cache entries are updated elsewhere. Cache-coherence controls bound staleness when the underlying source changes.

Although Bypass the cache for critical reads and use an appropriate strongly consistent data path can be valid when An application requires strict financial balance reads and a proposed eventually consistent cache may show obsolete values, it targets a different technical condition. Strong consistency requirements can outweigh the performance benefit of stale cached data.

This alternative, Review TTL, key distribution and read-through cache population, is appropriate when An ElastiCache deployment shows a high miss rate and database traffic did not fall. Cache effectiveness depends on repeated keys and suitable expiration rules.

For the separate situation A global media site retrieves the same files from origin for most requests despite a CloudFront distribution, Inspect cache-key headers, cookies and TTL policies to improve cacheability could be useful. Excessive cache-key variation fragments edge cache hit rates.

This alternative, Review TTL, key distribution and read-through cache population, is appropriate when An ElastiCache deployment shows a high miss rate and database traffic did not fall. Cache effectiveness depends on repeated keys and suitable expiration rules.

For the separate situation A database cache contains all products but frequently accessed records still miss because cache memory is undersized, Right-size the cache working set and inspect evictions and memory pressure could be useful. Capacity pressure can evict hot keys before they are reused.

Bypass the cache for critical reads and use an appropriate strongly consistent data path solves the related scenario An application requires strict financial balance reads and a proposed eventually consistent cache may show obsolete values; strong consistency requirements can outweigh the performance benefit of stale cached data.

This alternative, Inspect cache-key headers, cookies and TTL policies to improve cacheability, is appropriate when A global media site retrieves the same files from origin for most requests despite a CloudFront distribution. Excessive cache-key variation fragments edge cache hit rates.

For the separate situation An application requires strict financial balance reads and a proposed eventually consistent cache may show obsolete values, Bypass the cache for critical reads and use an appropriate strongly consistent data path could be useful. Strong consistency requirements can outweigh the performance benefit of stale cached data.

Add explicit invalidation or short TTL for critical mutable inventory fields solves the related scenario A product API returns stale inventory levels after cache entries are updated elsewhere; cache-coherence controls bound staleness when the underlying source changes.

For the separate situation A global media site retrieves the same files from origin for most requests despite a CloudFront distribution, Inspect cache-key headers, cookies and TTL policies to improve cacheability could be useful. Excessive cache-key variation fragments edge cache hit rates.

Review TTL, key distribution and read-through cache population solves the related scenario An ElastiCache deployment shows a high miss rate and database traffic did not fall; cache effectiveness depends on repeated keys and suitable expiration rules.

Although Add explicit invalidation or short TTL for critical mutable inventory fields can be valid when A product API returns stale inventory levels after cache entries are updated elsewhere, it targets a different technical condition. Cache-coherence controls bound staleness when the underlying source changes.