Core Services and the Well-Architected Lens
Module 2 · Technical Breadth

Core Services and the Well-Architected Lens

The breadth of AWS knowledge a TAM interview expects (compute, storage, networking, databases, security, and observability) organized around the six Well-Architected pillars and the resilience patterns customers ask about most.

Breadth Over Depth (But Some Depth)

A TAM supports customers who use dozens of services. Interviewers usually probe breadth, then pick one area you claim as a strength and go deep to see where your knowledge ends. A common shape:

  • “Walk me through what happens when a user types our URL and gets a page.” (networking, DNS, load balancing, compute)
  • “How would you make this architecture survive the loss of an Availability Zone?” (resilience)
  • “Tell me about the deepest technical problem you’ve solved.” (depth, and the Dive Deep Leadership Principle)

An AWS Solutions Architect – Associate certification is a good benchmark for breadth. Many postings list AWS certifications as preferred, not required.


The Service Map You Should Know Cold

AreaKnow wellKnow what it’s for
ComputeEC2 (instance families, Auto Scaling, purchase options), Lambda, ECS/EKS, FargateBatch, Elastic Beanstalk, App Runner
StorageS3 (storage classes, lifecycle, versioning), EBS (gp3, io2, snapshots), EFSFSx family, Storage Gateway, Backup
NetworkingVPC, subnets, route tables, security groups vs NACLs, NAT, ELB (ALB/NLB), Route 53, CloudFrontTransit Gateway, Direct Connect, PrivateLink, Global Accelerator
DatabasesRDS and Aurora (Multi-AZ, read replicas, backups), DynamoDB (capacity modes, partition keys)ElastiCache, Redshift, OpenSearch, DocumentDB
SecurityIAM (roles, policies, least privilege), KMS, Organizations and SCPs, CloudTrailGuardDuty, Security Hub, WAF, Shield, Secrets Manager
ObservabilityCloudWatch (metrics, logs, alarms), CloudTrail, VPC Flow LogsX-Ray, Config, Systems Manager
IntegrationSQS, SNS, EventBridgeStep Functions, Kinesis, MSK

You don’t need every feature. You need to explain what each service is for, how it fails, and how a customer would know.


The Six Well-Architected Pillars

The AWS Well-Architected Framework gives TAMs a shared vocabulary with customers. Know all six pillars and one or two practices for each:

PillarCore questionExample practices
Operational ExcellenceCan you run and improve this workload?Infrastructure as code, runbooks, small reversible changes, post-incident learning
SecurityAre data and systems protected?Least privilege, encryption, traceability, automated detection
ReliabilityDoes it recover and meet demand?Multi-AZ, health checks, auto-recovery, tested backups, quota management
Performance EfficiencyAre resources used efficiently as needs change?Right instance types, caching, serverless, measure before optimizing
Cost OptimizationAre you paying only for what you need?Right-sizing, Savings Plans, lifecycle policies, cost allocation tags
SustainabilityAre you minimizing environmental impact?Higher utilization, efficient instance types (like Graviton), data lifecycle

A Well-Architected review walks a workload through these questions, records high- and medium-risk issues, and produces an improvement plan. TAMs often run these reviews or bring in an SA to run them. The TAM then tracks the improvement plan to done.

⚠️ Common Pitfall: Treating the Review as the Outcome

Common mistake: a candidate proudly describes running 20 Well-Architected reviews.

The interviewer’s follow-up will be: how many high-risk issues got fixed? A review that produces a list nobody acts on has no impact. Talk about remediation: how you prioritized the risks, got owners and dates, and followed up.


Resilience Patterns Customers Ask About

Know these terms and be ready to whiteboard them:

  • Availability Zones vs Regions. Multi-AZ protects against a data-center-level failure. Multi-Region protects against a regional event, at much higher cost and complexity.
  • RTO and RPO. Recovery time objective is how long you can be down. Recovery point objective is how much data you can lose.
  • DR strategies, from cheapest and slowest to most expensive and fastest: backup and restore → pilot light → warm standby → multi-site active/active.
  • Static stability. Design so the system keeps working when a dependency fails, without needing to make changes during the failure (for example, pre-provisioning capacity across AZs instead of relying on scaling up during an outage).
  • Blast radius. Use cells, separate accounts, and shuffle sharding to limit how much one failure affects.
  • Quotas. Many launch-day outages are actually service quota limits nobody raised in time.

The four DR strategies trade cost against recovery speed:

Backup and restoreRecovery: hoursBackups onlyPilot lightRecovery: 10s of minData live, servers offWarm standbyRecovery: minutesSmall copy runningActive/activeRecovery: near zeroAll Regions liveHigher cost and complexityFaster recovery (RTO) and less data loss (RPO)

Key Takeaway

Interviewers test breadth across compute, storage, networking, databases, security, and observability, then go deep in one area. Frame your answers with the six Well-Architected pillars, know RTO/RPO and the four DR strategies, and remember that a review counts only when its risks get fixed.

🧪 Try It: Whiteboard Drill: Survive an AZ Failure

A customer runs a web app on two EC2 instances in one AZ behind an ALB, with a single-AZ RDS MySQL database and user uploads saved on the instances’ local EBS volumes.

List every single point of failure, then redesign it to survive the loss of an AZ. Name the services and settings you’d change.

A strong answer covers: an Auto Scaling group spanning at least two AZs (with enough capacity in each to carry the load alone), ALB subnets in multiple AZs, RDS Multi-AZ (or Aurora), uploads moved to S3, health checks, and a backup and restore test. It also mentions the cost of the extra capacity and how to explain it to the customer.

🧠 Knowledge Check

1. Which Well-Architected pillar most directly covers Multi-AZ deployments and tested backups?

2. A customer can tolerate losing at most 5 minutes of data. Which metric is that?

3. Which DR strategy keeps a scaled-down but fully functional copy of production running in another Region?

4. What should a TAM emphasize when describing Well-Architected reviews in an interview?