AWS Solutions Architect Professional flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

RPO versus RTO

Recovery point objective is the maximum acceptable data loss measured in time; recovery time objective is the maximum acceptable downtime before the workload must be restored and serving users again.

Aurora replica lag

Aurora replicas read from the same distributed storage volume as the writer, so replica lag is typically under 100 milliseconds and a failed writer is promoted in about 30 seconds.

SQS visibility timeout

The period during which a received message is hidden from other consumers. If the consumer does not delete it before the timeout expires, the message becomes visible and is delivered again.

Dead-letter queue maxReceiveCount

A redrive policy moves a message to a dead-letter queue after it has been received maxReceiveCount times without being deleted, isolating poison messages from the main queue.

Spot Instance interruption notice

AWS delivers a two-minute warning through instance metadata and EventBridge before reclaiming a Spot Instance, so workloads must checkpoint, drain, or terminate within that window.

Route 53 failover routing

Sends traffic to a primary record while an associated Route 53 health check passes and to a secondary record when it fails; health checks can be endpoint, calculated, or CloudWatch-alarm based.

Multi-AZ standby versus read replica

A Multi-AZ RDS standby is synchronously replicated and not readable, serving only automatic failover; a read replica is asynchronously replicated and readable, and must be promoted manually.

ALB versus NLB

ALB works at layer 7 with host, path, and header routing for HTTP and HTTPS; NLB works at layer 4, preserves source IP, offers static IPs, and handles extreme throughput with very low latency.

Lambda invocation timeout limit

A Lambda function can run at most 15 minutes per invocation and use up to 10,240 MB of memory; longer-running workloads need ECS, EKS, or AWS Batch instead.

EFS versus FSx

EFS provides shared elastic NFS storage for Linux workloads; FSx for Windows File Server provides SMB with Active Directory integration, and FSx for Lustre delivers high-throughput parallel file systems for HPC.

DynamoDB global tables

Global tables replicate a DynamoDB table across Regions in an active-active configuration using last-writer-wins conflict resolution, giving multi-Region reads and writes with low local latency.

Kinesis versus SQS

Kinesis Data Streams retains an ordered record stream that multiple consumers can replay within the retention window; SQS deletes each message after acknowledgement and supports only competing consumers.

Athena on S3

Athena is a serverless query service that runs standard SQL directly against data in S3 and charges per terabyte scanned, so partitioning and columnar formats drive both cost and speed.

AWS PrivateLink

PrivateLink exposes a service to consumer VPCs through interface endpoints, keeping traffic on the AWS network with no VPC peering, route table changes, or CIDR overlap risk.

Storage Gateway File Gateway

File Gateway presents NFS or SMB file shares backed by S3 objects, caching frequently accessed data locally while retaining a full durable copy in S3 for lifecycle and archive policies.

AWS Application Migration Service

MGN replicates source servers block-by-block into AWS with an installed agent, lets you launch non-disruptive test instances, and performs the final cutover with minimal downtime.

DMS and SCT

Schema Conversion Tool rewrites incompatible schema and code for heterogeneous migrations; Database Migration Service then moves the data, optionally with change data capture for near-zero-downtime cutover.

Snowball Edge use case

Choose Snowball Edge when moving many terabytes would exceed available network bandwidth or take longer than about a week; data is encrypted at rest and shipped to an AWS Region.

The 6 Rs of migration

Migration strategies: rehost (lift and shift), replatform (lift, tinker, shift), refactor, repurchase, relocate, retain, and retire; retiring an application is often the cheapest outcome.

Route 53 weighted cutover

Weighted DNS records gradually shift traffic from on-premises to AWS during a migration cutover and allow rollback by adjusting weights; keep TTLs low before the migration window.

SCP permission behavior

Service control policies set the maximum permissions available to accounts in an organization or OU but grant nothing themselves; they do not affect the management account or service-linked roles.

CloudFormation StackSets

StackSets deploy a single template to many accounts and Regions from one operation, using service-managed permissions through AWS Organizations or self-managed IAM roles in target accounts.

IAM permissions boundary

A permissions boundary caps the maximum permissions that identity-based policies can grant to an IAM user or role; effective permissions are the intersection of the policy and the boundary.

Centralized CloudTrail organization trail

An organization trail delivers events from every account to one S3 bucket in a log-archive account, while SCPs stop member accounts from disabling or altering the trail.

VPC sharing with RAM

AWS RAM shares subnets from an owner VPC into participant accounts, which launch resources into those subnets while the owner keeps control of routing, network ACLs, and CIDR allocation.

IAM Identity Center

Centrally manages workforce identities, federates with an external identity provider, and assigns permission sets to accounts in AWS Organizations, replacing per-account IAM users and passwords.

Savings Plans versus Reserved Instances

Compute Savings Plans cover EC2, Fargate, and Lambda across families and Regions for up to 66% off; EC2 Instance Savings Plans trade that flexibility for discounts up to 72%.

S3 Intelligent-Tiering

Automatically moves objects between frequent and infrequent access tiers as access patterns change, charging a small per-object monitoring fee and no retrieval fees for tier transitions.

gp3 versus gp2

gp3 decouples provisioned IOPS and throughput from volume size, starting at 3,000 IOPS and 125 MB/s at any capacity; gp2 performance scales with size at 3 IOPS per GB.

NAT Gateway cost

A NAT Gateway is billed per hour plus per gigabyte processed; gateway VPC endpoints for S3 and DynamoDB keep that traffic private and avoid the NAT data processing charge entirely.

Cost allocation tags

Tags must be activated as cost allocation tags in the billing console before they appear in Cost Explorer and the Cost and Usage Report; activation can take up to 24 hours.

AWS Compute Optimizer

Analyzes CloudWatch metrics to recommend right-sizing for EC2 instances, Auto Scaling groups, EBS volumes, Lambda functions, and Fargate tasks, flagging over-provisioned resources.