Well-Architected Tool reviews: Security pillar, Financial Services Industry Lens and tracking improvements
Run a structured Well-Architected review of a Harbourline workload in the AWS Well-Architected Tool, apply the Financial Services Industry Lens, turn high-risk findings into an owned improvement plan, and track progress with milestones and a C# risk report.
Assumes: .NET 10 LTS · AWS SDK for .NET v4 (AWSSDK.WellArchitected 4.x) · AWS CLI v2 · CDK v2 (Amazon.CDK.Lib 2.277.0) · Terraform AWS provider v6 · Region us-east-1. Verified against official sources on 2026-11-18.
Learning objectives
By the end of this lesson you can:
- Describe the six Well-Architected pillars and what the AWS Well-Architected Tool adds to the framework document.
- Define a workload, apply the Well-Architected Framework lens and the Financial Services Industry Lens, and answer questions with evidence notes.
- Interpret high-risk and medium-risk issues, and turn them into an improvement plan with owners and dates.
- Save milestones so progress between reviews is measurable.
- Run a review as a short facilitated workshop with the right people in the room.
- Pull risk counts and stale reviews from the Well-Architected Tool API with a C# tool.
Prerequisites
- Well-Architected framework introduction for the six pillars and design principles. This lesson is about using the Tool in a governed way.
- Module 5 security overview, since the Security pillar review draws on identity, detection, data protection and incident response.
- FinOps foundations and Commitments and rightsizing for the Cost Optimization pillar.
- Tools: AWS CLI v2 with
learn-sandbox, .NET 10 SDK. A workload to review: use the Harbourline Accounts API from earlier modules, or the design on paper.
Why this matters in banking
Reviews are a normal part of change control in a bank. A new payment product goes through architecture review, security review, operational resilience review and sometimes a regulatory notification. Each reviewer has their own checklist and their own spreadsheet. A Well-Architected review gives you a shared, vendor-neutral set of questions about operations, security, reliability, performance, cost and sustainability, and an output you can store: answers with notes, a list of high-risk items and an improvement plan.
It is not an audit and it is not a certificate. It is a structured conversation that makes risk visible early. A risk board can ask "How many high-risk items does the Payments platform have, and how has that changed since last quarter?" and get an answer from milestones rather than from memory. The Financial Services Industry Lens adds questions that speak to the concerns of regulated firms: operational resilience, data privacy, auditability, third-party risk, and controls for new service adoption.
A review result is not an assurance or compliance statement
Answering "yes" to a question in the Tool means that the team believes it follows a best practice. It does not mean the control has been tested, it does not map one-to-one to a regulation, and it does not show that you comply with the PRA, FCA, EBA, PCI DSS or UK GDPR requirements. Compliance and assurance remain with your risk, compliance and internal audit functions. Use the Tool's output as an input to them, not a replacement.
The idea in plain language
The AWS Well-Architected Framework is a set of questions and best practices grouped into six pillars:
| Pillar | The question in one line |
|---|---|
| Operational excellence | Can we run it, see it and improve it? |
| Security | Is it protected, and can we prove who did what? |
| Reliability | Does it recover from failure and keep serving? |
| Performance efficiency | Do we use the right resources for the load? |
| Cost optimization | Are we spending only where it creates value? |
| Sustainability | Are we minimising the environmental impact? |
The AWS Well-Architected Tool is a free service in the console (and API) where you record a review. You define a workload (the set of resources and code that delivers a business outcome), choose one or more lenses (question sets), answer the questions, and the Tool shows risks: high risk issues (HRI) and medium risk issues (MRI). You then work through an improvement plan, save a milestone (a snapshot) and review again later.
Analogy: a Well-Architected review is a pre-flight checklist read aloud by two people, with a record of what was checked and what was waved through. It does not fly the plane, but it is hard to claim that nobody knew the fuel gauge was broken.
Lenses relevant to Harbourline
| Lens | What it adds |
|---|---|
| AWS Well-Architected Framework | The base set of questions across six pillars |
| Financial Services Industry Lens | Best practices and questions for financial institutions: resiliency, security, data privacy, governance and cost. It is available as an AWS-official lens in the Tool's lens catalogue. The published document is dated 27 January 2026 |
| Serverless, Data Analytics, Machine Learning, SaaS and others | Domain lenses for specific architectures. ⚠ VERIFY the current list in the lens catalogue |
| Custom lens | Your own questions: for example Harbourline's internal control standards, written as a lens |
The Financial Services Industry Lens is not a regulation mapped to controls. In its own introduction, it describes guidance for designing workloads that align with the risk and control objectives you define, including those that help you meet the expectations of supervisors. Treat that as a design aid.
What problem the AWS service solves
| Need | Tool feature | Limit |
|---|---|---|
| Consistent review questions | Lenses | Questions are generic; they do not know your architecture |
| Evidence of what was decided | Notes on each answer, and milestones | Notes are only as good as the people writing them |
| Prioritising fixes | Risk ratings and improvement plan with links to guidance | Ratings derive from your answers; honest answers matter |
| Tracking over time | Milestones, risk counts and improvement status | A milestone is a snapshot; nothing forces you to take one |
| Sharing | Share a workload or custom lens with other accounts or the organisation (via AWS Resource Access Manager) | Needs sharing configured and permissions set |
| Your own standards | Custom lenses (JSON) | You write and maintain them |
| Cross-account reviews | Workload can reference multiple account IDs | The Tool stores answers, not live telemetry |
The Tool itself has no charge (you pay only for any AWS resources you use). ⚠ VERIFY on the current pricing page.
Honest limits. The Tool does not scan your account to answer for you. (There are integrations with Trusted Advisor and AWS Config-based checks to pre-populate some answers in some pillars, ⚠ VERIFY current integrations.) It relies on people. A review where one engineer ticks the boxes alone is close to useless, and a review run as a blame exercise produces dishonest answers.
Microsoft / Azure comparison
| You know (Microsoft/Azure) | AWS equivalent | What's different |
|---|---|---|
| Azure Well-Architected Review (assessment) | AWS Well-Architected Tool | Both are question-based; AWS has six pillars including sustainability, Azure has five |
| Azure Advisor score | Not the same: AWS gives per-question risk, not a single score | You track HRI/MRI counts and milestones instead |
| Azure Architecture Center workload guidance | Well-Architected lenses | AWS publishes lenses per workload type and industry |
| Microsoft Cloud Adoption Framework, landing zone reviews | Well-Architected Tool plus Control Tower and Security Reference Architecture | The Tool reviews a workload; landing-zone governance is a different layer |
| Azure Policy compliance assessments | Config conformance packs (see the evidence lesson) | Policy and Config test configuration automatically; the Tool captures human judgement |
| Review templates and Azure DevOps boards | Review templates, profiles and improvement plan links | Link improvement items to your own ticket system |
Architecture
Security-boundary view: the Tool stores answers and notes in the account where the workload is defined. Notes can contain architecture details and known weaknesses. Treat a workload review as a confidential document: restrict who can read it (IAM and sharing permissions), keep it out of tickets that many people can see, and do not paste secrets or customer data into notes.
Step-by-step implementation
Choose the workload and scope it
Pick one business outcome, for example "Accounts API production". Write down its accounts, Regions, data classification, dependencies and owner. Reviewing "the whole bank" is not a workload.
Prepare the evidence
Before the workshop, collect the architecture diagram, the last penetration test summary, the recovery test results, the monthly cost report and the incident list. The best review answers can point to something.
Create the workload and choose lenses
Create the workload in the Tool as Production (or Pre-production for a design review), with the Framework lens and the Financial Services Industry Lens. Start with the Security pillar if time is short.
Run the workshop
Two to three hours per lens is realistic. Read each question, discuss, choose the practices you really follow and write a note with evidence or the reason for a gap. Mark questions that do not apply as not applicable with a reason.
Read the risks
Look at HRI first. Ask for each: is it real, what is the impact, what is the smallest fix? Resist arguing the rating down. Record disagreement in the notes.
Build the improvement plan
For every HRI, and the MRIs you choose, create a ticket in your own system with an owner, a date and a link to the question. Add a link back from the ticket to the workload.
Save a milestone
Save Milestone 1: baseline. After 90 days, review again and save Milestone 2. The change in risk counts is your progress story.
AWS Console instructions
- Open the Tool: Console → search Well-Architected Tool (or open it from Management and Governance).
- Create a workload: Define workload → enter Name, Description, Review owner, Environment (Production or Pre-production), Regions and optionally account IDs, industry type and architectural design link. Choose Next.
- Apply lenses: On the lens step, tick AWS Well-Architected Framework and add the Financial Services Industry Lens from the lens catalogue (select it from the list of AWS official lenses). Choose Define workload.
- Answer questions: Workloads → your workload → Start reviewing. Choose a lens, pick a pillar, answer questions by selecting the best practices you follow, and use Notes on each.
- Improvement plan: Workloads → your workload → Improvement plan shows items by risk. Each item links to guidance and a "helpful resource".
- Milestone: Milestones → Save milestone, give it a name such as
2026-Q4-baseline. - Share: Workloads → your workload → Sharing → Share with an account or organisation, choosing read-only or contributor permission. ⚠ VERIFY the exact wording.
- Report: Generate report produces a PDF of the answers and improvement plan. Check what is in it before you send it anywhere.
Console wording changes. ⚠ VERIFY each path.
AWS CLI
# 1) Which lenses can I use? (look for the FSI lens and note its alias)
aws wellarchitected list-lenses --profile learn-sandbox --region us-east-1 \
--query 'LensSummaries[].[LensName,LensAlias,LensVersion]' --output table
# 2) Create a workload with the Framework lens. The FSI lens can be associated next.
aws wellarchitected create-workload \
--workload-name "hl-accounts-api-prod" \
--description "Harbourline Accounts API, customer-facing, UK" \
--environment PRODUCTION \
--aws-regions us-east-1 \
--review-owner "payments-platform-team@example.com" \
--lenses wellarchitected \
--tags Project=cea-lab,Lab=m15-l12,Environment=learn \
--profile learn-sandbox --region us-east-1
# 3) Associate the Financial Services Industry Lens (replace the alias with the one from step 1).
aws wellarchitected associate-lenses --workload-id <workload-id> \
--lens-aliases <fsi-lens-alias> --profile learn-sandbox --region us-east-1
# 4) See the questions in the security pillar and their IDs.
aws wellarchitected list-answers --workload-id <workload-id> \
--lens-alias wellarchitected --pillar-id security \
--profile learn-sandbox --region us-east-1
# 5) Answer a question: choose the best practices you follow and record notes.
aws wellarchitected update-answer --workload-id <workload-id> \
--lens-alias wellarchitected --question-id <question-id> \
--selected-choices <choice-id-1> <choice-id-2> \
--notes "Roles via IAM Identity Center; break-glass tested 2026-09. Evidence: ticket SEC-1142." \
--profile learn-sandbox --region us-east-1
# 6) Review summary and improvement items for the framework lens.
aws wellarchitected get-lens-review --workload-id <workload-id> --lens-alias wellarchitected \
--profile learn-sandbox --region us-east-1
aws wellarchitected list-lens-review-improvements --workload-id <workload-id> \
--lens-alias wellarchitected --pillar-id security \
--profile learn-sandbox --region us-east-1
# 7) Save a milestone, then list milestones.
aws wellarchitected create-milestone --workload-id <workload-id> --milestone-name "2026-Q4-baseline" \
--profile learn-sandbox --region us-east-1
aws wellarchitected list-milestones --workload-id <workload-id> --profile learn-sandbox --region us-east-1
# 8) Share with another account (read-only).
aws wellarchitected create-workload-share --workload-id <workload-id> \
--shared-with 222233334444 --permission-type READONLY \
--profile learn-sandbox --region us-east-1Question IDs and choice IDs come from list-answers and get-answer, and differ between lenses and lens versions. Never copy IDs from a blog post. ⚠ VERIFY every flag with aws wellarchitected <command> help.
A small custom lens
A custom lens lets you encode Harbourline's own control standard as questions. The schema below follows the documented custom lens JSON format. ⚠ VERIFY the schema version and required fields against the current documentation before importing.
{
"schemaVersion": "2021-11-01",
"name": "Harbourline Controls Lens",
"description": "Internal control questions that every customer-facing workload must answer.",
"pillars": [
{
"id": "hlcontrols",
"name": "Harbourline Controls",
"questions": [
{
"id": "tagging",
"title": "How do you identify the owner and cost centre of every resource?",
"description": "Resources must be attributable to a team and a cost centre.",
"helpfulResourceDisplayText": "Harbourline tag standard",
"choices": [
{
"id": "tags-from-code",
"title": "All resources receive the mandatory tags from infrastructure as code."
},
{
"id": "tag-policy",
"title": "A tag policy and a detective rule check tag coverage."
},
{
"id": "none",
"title": "None of these."
}
],
"riskRules": [
{ "condition": "tags-from-code && tag-policy", "risk": "NO_RISK" },
{ "condition": "tags-from-code || tag-policy", "risk": "MEDIUM_RISK" },
{ "condition": "default", "risk": "HIGH_RISK" }
]
}
]
}
]
}Import with aws wellarchitected import-lens --json-string "$(cat harbourline-controls-lens.json)" --lens-alias hl-controls (check parameter names with help), then publish a version with create-lens-version before using it on workloads. ⚠ VERIFY.
Infrastructure as Code
A workload definition as code is useful for governance: every new production workload is registered the same way, with the same lenses and an owner. The Tool's questions and answers still belong to people, so only the workload shell is created. Property names follow CloudFormation AWS::WellArchitected::Workload; the code below compiles against Amazon.CDK.Lib 2.277.0. Pillar priorities are not set here: add them in the console or with update-workload. ⚠ VERIFY property availability if you use another version.
using Amazon.CDK;
using Amazon.CDK.AWS.WellArchitected;
using Constructs;
namespace Harbourline.Governance.Infra;
public sealed class WellArchitectedWorkloadStack : Stack
{
public WellArchitectedWorkloadStack(Construct scope, string id, IStackProps? props = null) : base(scope, id, props)
{
Amazon.CDK.Tags.Of(this).Add("Project", "cea-lab");
Amazon.CDK.Tags.Of(this).Add("Lab", "m15-l12");
Amazon.CDK.Tags.Of(this).Add("Environment", "learn");
new CfnWorkload(this, "AccountsApiWorkload", new CfnWorkloadProps
{
WorkloadName = "hl-accounts-api-prod",
Description = "Harbourline Accounts API, customer-facing, UK",
Environment = "PRODUCTION",
ReviewOwner = "payments-platform-team@example.com",
Lenses = new[] { "wellarchitected" }, // add the FSI lens alias once you have looked it up
AwsRegions = new[] { "us-east-1" },
});
}
}Complete code
A C# report for a risk board. For every workload it prints the high and medium risk counts, how many questions remain unanswered, the date of the last milestone, and a flag when a workload has high risks and no milestone in the last 90 days.
<Project Sdk="Microsoft.NET.Sdk">
<PropertyGroup>
<OutputType>Exe</OutputType>
<TargetFramework>net10.0</TargetFramework>
<Nullable>enable</Nullable>
<ImplicitUsings>enable</ImplicitUsings>
<TreatWarningsAsErrors>true</TreatWarningsAsErrors>
</PropertyGroup>
<ItemGroup>
<!-- Look up exact 4.x versions at https://api.nuget.org/v3-flatcontainer/<id>/index.json -->
<PackageReference Include="AWSSDK.WellArchitected" Version="4.*" />
</ItemGroup>
</Project>using Amazon.WellArchitected;
using Amazon.WellArchitected.Model;
const int staleDays = 90;
var now = DateTime.UtcNow;
using var wa = new AmazonWellArchitectedClient();
var summaries = new List<WorkloadSummary>();
string? token = null;
do
{
var page = await wa.ListWorkloadsAsync(new ListWorkloadsRequest { MaxResults = 50, NextToken = token });
summaries.AddRange(page.WorkloadSummaries ?? []);
token = string.IsNullOrEmpty(page.NextToken) ? null : page.NextToken;
}
while (token is not null);
Console.WriteLine($"{"Workload",-34} {"HRI",4} {"MRI",4} {"Unanswered",10} {"Last milestone",-15} Flag");
int flagged = 0;
foreach (var w in summaries.OrderBy(s => s.WorkloadName, StringComparer.OrdinalIgnoreCase))
{
var counts = w.RiskCounts ?? new Dictionary<string, int>();
int high = counts.GetValueOrDefault("HIGH");
int medium = counts.GetValueOrDefault("MEDIUM");
int unanswered = counts.GetValueOrDefault("UNANSWERED");
// Latest milestone date for this workload.
var milestones = await wa.ListMilestonesAsync(new ListMilestonesRequest { WorkloadId = w.WorkloadId, MaxResults = 50 });
DateTime? last = (milestones.MilestoneSummaries ?? [])
.Select(m => (DateTime?)m.RecordedAt)
.OrderByDescending(d => d)
.FirstOrDefault();
bool stale = last is null || (now - last.Value).TotalDays > staleDays;
string flag = high > 0 && stale ? "HIGH RISK + STALE" : high > 0 ? "high risk" : stale ? "stale" : "";
if (flag.Length > 0) flagged++;
Console.WriteLine($"{w.WorkloadName,-34} {high,4} {medium,4} {unanswered,10} {(last?.ToString("yyyy-MM-dd") ?? "never"),-15} {flag}");
}
Console.WriteLine($"\n{summaries.Count} workloads, {flagged} need attention.");
Console.WriteLine("Counts reflect what teams recorded. They are not assurance that controls work.");Code explained line by line
staleDaysis a policy constant. A review older than a quarter, with no new milestone, is treated as stale. Choose a cadence with your risk function.ListWorkloadsAsyncreturns summaries in pages. The loop followsNextTokenuntil it is empty.RiskCountsis a dictionary keyed by risk name (HIGH,MEDIUM,NONE,UNANSWERED,NOT_APPLICABLE).GetValueOrDefaulthandles a missing key as zero. ⚠ VERIFY the exact key names against a real response.- A second call per workload fetches milestones. For a very large number of workloads, cache results or run the report nightly rather than on demand.
- The latest
RecordedAtis the date of the last snapshot.nullmeans the team never saved one, which is itself a finding. - The flag combines the two things a risk board cares about: open high risks, and not looking at them recently.
- The tool is read-only and needs
wellarchitected:ListWorkloadsandwellarchitected:ListMilestones(andGetWorkloadif you extend it). - The closing line is deliberate. Summaries of review data are easy to over-read as assurance. Say what they are.
Security considerations
⚠ Learning shortcut — not for production
The lab reviews a single workload alone, in your own sandbox account, and stores honest notes in plain text. In production, reviews are held in a dedicated governance account with controlled sharing, notes avoid secrets and customer data, the review is facilitated with the right owners, high risks have tickets and due dates, and the risk register links to the Tool. The reporting role is read-only, and report PDFs are classified and stored like other security documents.
- Reviews describe weaknesses. A list of high-risk issues is a roadmap for an attacker. Restrict access and use sharing permissions deliberately.
- No secrets, no personal data. Notes should reference a ticket or a document, not paste a configuration.
- Do not game the rating. Pressure to remove high risks by choosing generous answers defeats the purpose. Make "we accept this risk" an explicit, signed decision in the risk register.
- Share least privilege. Prefer read-only sharing, and contributor only for the review team.
- Retention. Milestones are historical records of what a team knew at a date. Keep them for as long as your records policy requires.
Banking use case
Harbourline's Cards platform is about to enter a new Region for its EU subsidiary. The architecture board requires a Well-Architected review with the Framework lens (Security, Reliability and Cost Optimization first) and the Financial Services Industry Lens. In a three-hour workshop with the architect, security lead, SRE, finance partner and a risk representative, the team records 6 high risks: one around third-party dependency exit planning, one around untested restoration of a data store, and four around incident response runbooks. Each becomes a ticket with an owner and a date. The baseline milestone goes in the board pack. Three months later the second milestone shows 2 high risks left, with one formally accepted and signed by the head of risk. The board sees progress, the auditors see a trail, and nobody mistakes the result for a compliance certificate. The FSI Lens questions on operational resilience feed into the separate operational resilience mapping that the compliance team owns.
Common mistakes
- Doing the review alone. One person guesses. Bring architect, security, operations, finance and risk.
- Reviewing too much at once. A whole organisation is not a workload. One business outcome per review.
- No evidence in notes. Without a pointer to proof, "yes" is just an opinion. Link a ticket or document.
- Arguing the rating down. If a risk is uncomfortable, fix it or formally accept it. Do not reword it away.
- No follow-up. Findings without owners and dates are a list of regrets.
- Treating the result as compliance. It is a design aid. State that on the cover of any report.
- Reviewing once. Architecture drifts. Re-review at least annually or after major change, with milestones in between.
- Ignoring Cost Optimization and Sustainability. They are often skipped because they are less urgent, and then the bill or the carbon target surprises everyone.
Troubleshooting
| Symptom | Likely cause | Fix |
|---|---|---|
| FSI lens not in the list | Wrong Region, or lens catalogue not browsed | List lenses with list-lenses or browse the catalogue in the console; check the Region |
ValidationException on update-answer | Wrong question or choice ID for that lens version | Fetch IDs with list-answers / get-answer for the same lens alias |
Workload shows UNANSWERED for many questions | Review not completed | Schedule a workshop; mark not-applicable with a reason |
| Cannot share the workload | Permissions or organisation sharing not set up | Check IAM permissions and RAM sharing settings with your platform team ⚠ VERIFY |
| Risk counts do not change after answering | Answer not saved, or a different lens reviewed | Save and refresh; check the lens alias |
AccessDeniedException from the C# report | Role lacks wellarchitected:List* | Add read-only permissions for the Tool |
| Custom lens import fails | JSON schema or ID mistakes | Validate the JSON, keep IDs lowercase and unique, see the schema documentation ⚠ VERIFY |
| Improvement plan has no links | Custom lens without helpful resources | Add helpfulResource fields in the lens |
Cost and cleanup
Cost: free. The Tool has no charge. ⚠ VERIFY the current pricing page. The workshop time is the real cost, and the main investment is people's time. There are no AWS resources to leave running.
Cleanup (only for the lab workload; real workloads should be retained):
# Delete the lab workload (this deletes its answers and milestones).
aws wellarchitected delete-workload --workload-id <workload-id> \
--client-request-token "$(uuidgen)" --profile learn-sandbox --region us-east-1
# If you imported a custom lens for practice, delete its shares and the lens.
aws wellarchitected delete-lens --lens-alias hl-controls --lens-status ALL \
--client-request-token "$(uuidgen)" --profile learn-sandbox --region us-east-1
# Remove local files.
rm -f harbourline-controls-lens.jsonCheck parameter names with help; delete calls require a client request token. ⚠ VERIFY. Deleting a workload is permanent, so export the PDF report first if you want to keep your notes.
Exercises
- 🟢 Guided: create a workload for the Accounts API (or a paper design), add the Framework lens, and answer all Security pillar questions with honest notes. Save a baseline milestone.
- 🟡 Challenge: add the Financial Services Industry Lens and answer five questions on resilience and third-party risk. Turn every high risk into a ticket draft with owner, due date and the smallest fix.
- 🔴 Live problem: your CTO asks for the Payments platform to be "Well-Architected certified" before a board meeting on Friday. Write a short reply: what a review is and is not, what you can deliver by Friday, and what you need from others.
Assignment
Add governance/well-architected/ to your portfolio repository with the workload definition as code, the exported PDF or Markdown summary of your Security pillar review, the improvement plan as a table (risk, owner, date), and the C# report.
Solutions
A good answer set has a note on every question: which practices you chose, and what proves it (for example "MFA enforced through Identity Center; see policy SEC-12 and the access review of 2026-09-30"). Expect a handful of high risks in a first review of a typical workload. That is normal: the first milestone is a baseline, not a score to be embarrassed by. After saving it, the Milestones page shows the risk counts frozen at that moment, so a later review can be compared against it.
For each high risk, write three lines: what the risk is in your own words, the smallest change that removes it, and the owner. For example, "No tested restore of the accounts data store" becomes "Schedule a quarterly restore test into an isolated account; owner: SRE lead; date: end of next month; evidence: test report in the evidence bucket." Group fixes by pillar and by owner, so that each team gets a short list, and flag any that need budget. Link the Financial Services Industry Lens questions to your resilience and third-party risk registers, owned by risk and compliance, rather than copying their content into the Tool.
Knowledge check
Answer: A saved snapshot of a workload's review state (answers, notes and risk counts) at a point in time. It lets you compare later reviews to the baseline.
Answer: No. The review records whether best practices were followed, according to the people answering. It is not an audit, it does not map to regulations, and compliance determinations belong to qualified assessors and your compliance function.
Answer: The Financial Services Industry Lens, available as an AWS-official lens in the Tool's lens catalogue, alongside the base Well-Architected Framework lens.
Answer: Without evidence a "yes" cannot be checked later. Notes that point to tickets, test reports or documents make the review credible and let a future reviewer see why an answer was given.
Answer: Create a ticket with an owner and due date (and the smallest fix), and either remediate it or record an explicit risk acceptance signed by the right authority. Then save a milestone and re-review.
Interview questions
- How would you run a Well-Architected review for a new payments service? Scope one workload, gather evidence, run a facilitated workshop with the framework and Financial Services Industry Lens, write notes with evidence, convert high risks to owned tickets, save a baseline milestone and re-review in a quarter.
- What is the difference between a Well-Architected review and an audit? A review is a collaborative self-assessment against best practices. An audit tests controls and evidence independently against a standard.
- How do you use a custom lens? To encode internal standards as questions, so that every workload answers them the same way, with risk rules that map answers to risk levels.
- What would you put on a risk board slide from the Tool? High and medium risk counts per workload, change since the last milestone, items overdue, and accepted risks.
Certification practice
A bank wants a repeatable way for every new customer-facing workload to be reviewed against both general best practices and financial-industry guidance, with results shared across accounts. Which approach BEST fits?
A. Spreadsheets maintained by each team. B. The AWS Well-Architected Tool with the Framework lens and the Financial Services Industry Lens, sharing workloads through AWS Resource Access Manager. C. AWS Config conformance packs only. D. AWS Trusted Advisor only.
Answer: B. The Tool provides a structured, repeatable review with lenses and sharing. A is inconsistent. C and D check configuration or generic best practices automatically but do not capture the architectural and process questions in the lenses.
Which Well-Architected pillar focuses on avoiding unnecessary costs and understanding spend?
A. Reliability B. Cost optimization C. Performance efficiency D. Operational excellence
Answer: B. The cost optimization pillar covers financial management, cost awareness and efficient use of resources. The others cover different concerns.
Which statement describes the AWS Well-Architected Tool?
A. A service that automatically fixes your architecture. B. A service in the console that helps you review workloads against best practices and track improvements. C. A firewall. D. A billing report.
Answer: B. The Tool guides a review and records improvements. It does not fix resources automatically, and it is neither a firewall nor a billing report.
Completion checklist
- I can name the six pillars and what each asks in one line
- I created a workload, applied two lenses and answered questions with evidence notes
- I saved a baseline milestone and turned high risks into owned tickets
- I can explain why a review is not compliance assurance
- I ran the C# report (or reviewed its output) and know the stale-review rule
- I ran the cleanup steps for the lab workload
Key takeaways
- The Tool records a structured human review. It does not inspect your account for you.
- Use the Framework lens plus the Financial Services Industry Lens for bank workloads, and a custom lens for your own standards.
- Evidence notes, owners, dates and milestones turn a workshop into a governance record.
- A review result is not an audit, a certificate or a statement of regulatory compliance.
- Re-review on a cadence, and report risk counts and trends to the risk board.
Further reading
Independent learning resource — not affiliated with, endorsed or sponsored by Amazon Web Services. AWS and related marks are trademarks of Amazon.com, Inc. or its affiliates. All banking data in examples is fictional. Sample architectures are for learning and are not automatically compliant with GDPR, PCI DSS or any banking regulation.
Commitments and rightsizing: Savings Plans, Reserved Instances, Compute Optimizer and Trusted Advisor
Cut a steady cloud bill safely: rightsize first, then size a Savings Plans commitment from your own usage data, understand Reserved Instances, and use Compute Optimizer and Trusted Advisor findings as a governed backlog.
Module 15 cheat sheet
One-page revision summary for Multi-Account Governance & FinOps: services, Azure mapping, CLI commands, limits, decision rules, exam traps and cost traps.