Overhead view of a laptop, an open notebook with a hand-drawn system diagram, headphones and a pen on a grey desk
    AWS Well-Architected

    AWS Well-Architected review and alignment

    A Well-Architected review is only useful if the risks it finds get fixed. Agnotic Technologies runs the review against the lenses that fit your workload, records it in the AWS Well-Architected Tool, ranks the high and medium risk issues and fixes them in milestones your team can review again.

    Lens choiceRecorded in the WA ToolRisk-ranked backlogMilestone re-reviews

    Trusted by global innovators

    Benchmark
    Chibasco
    Fundency
    Lantimer
    Lauren
    Lera
    One Minute
    Pento Pix
    TAP
    Xtrium
    Healthevolve
    Benchmark
    Chibasco
    Fundency
    Lantimer
    Lauren
    Lera
    One Minute
    Pento Pix
    TAP
    Xtrium
    Healthevolve
    Benchmark
    Chibasco
    Fundency
    Lantimer
    Lauren
    Lera
    One Minute
    Pento Pix
    TAP
    Xtrium
    Healthevolve
    Benchmark
    Chibasco
    Fundency
    Lantimer
    Lauren
    Lera
    One Minute
    Pento Pix
    TAP
    Xtrium
    Healthevolve

    Well-Architected

    What is an AWS Well-Architected review and alignment?

    An AWS Well-Architected review checks a workload against AWS best practice questions and lists its high and medium risk issues; alignment is fixing them. Agnotic Technologies picks the lenses that fit the workload, records the review in the Well-Architected Tool, ranks the risks into a backlog and remediates them in milestones you can re-review.

    The review is the start, not the deliverable. Most of the value is in the remediation that follows and in the milestone that shows it was done.

    A broader cloud assessment, landing zone design or a move to AWS sits with cloud consulting and AWS migration. Deep cost pillar work sits with cloud cost optimization.

    Lens choice

    Which Well-Architected lens fits your workload?

    The Framework lens applies to every workload. We add the AWS lens that matches what the workload does, because each one brings its own questions and improvement plan.

    Workload typeLens to addWhat it adds to the reviewEvidence a reviewer asks for
    Workload holding health dataHealthcare Industry LensHealthcare design principles and best practices across the six pillarsPHI data flow diagram, list of services covered by the AWS BAA, encryption settings and access logs
    Payments or bankingFinancial Services Industry LensFinancial services design principles and best practices across the six pillarsResilience and recovery tests, key management, audit trail and data residency decisions
    Multi-tenant SaaSSaaS LensTenant isolation, data partitioning, noisy neighbours, onboarding and tenant tiersThe isolation model, per-tenant usage and cost data and cross-tenant access tests
    Serverless API or event processingServerless Applications LensCommon serverless scenarios and the practices that keep them reliableFunction permissions, concurrency and timeout settings, retries and dead-letter handling
    Generative AI featureGenerative AI LensPractices for building generative AI workloads on AWSModel and prompt evaluation records, data access boundaries, guardrails and cost per request

    Custom lenses add your own questions, such as internal engineering standards, and run in the same Tool beside the AWS lenses.

    Review workflow

    How does a review run in the Well-Architected Tool?

    Recording the review in the Tool keeps the findings, the improvement plan and the milestones in one place your team can reopen.

    1.

    Define the workload

    Create the workload record with its accounts, Regions, environment and the lenses to apply, so findings stay tied to one system.

    Define
    2.

    Answer the questions

    Work through each lens with the engineers who built the system, checking answers against the accounts rather than memory.

    Review
    3.

    Record the risks

    The Tool flags high and medium risk issues from the answers; we add notes on the evidence behind each one.

    Record
    4.

    Save a baseline milestone

    Save the starting state as a milestone, so later progress is measured against it.

    Baseline
    5.

    Fix and re-review

    Work the backlog, update the answers and save a new milestone after each batch of fixes.

    Improve

    Remediation backlog

    How do we rank the risks the review finds?

    The Tool says which issues are high or medium risk. It does not say which to fix first, so we rank each one by risk and effort.

    01

    High risk, low effort

    Example: No MFA on the root user, or an S3 bucket open to the public

    Approach: Fixed first, often as one change to an organization-wide policy.

    02

    High risk, high effort

    Example: A single Availability Zone database behind a production workload

    Approach: Planned as its own piece of work with a tested failover and a maintenance window.

    03

    Medium risk, low effort

    Example: No alarms on queue depth or error rates

    Approach: Batched with other quick fixes, so each milestone closes a visible group of issues.

    04

    Medium risk, high effort

    Example: An older service with no automated deployment or rollback

    Approach: Scheduled against product work, or accepted with a written reason and a review date.

    Cloud audit

    What does a cloud audit for optimisation and performance turn into?

    Findings only matter once they become tasks. These are typical remediation tasks we take on after a review.

    Reliability

    • Multi-AZ databases
    • Backups with restore tests
    • Health checks and auto recovery
    • Service quotas watched before limits

    Performance efficiency

    • Instance and storage right-sizing
    • Caching for hot reads
    • Load tests before scaling changes
    • Newer instance families

    Security

    • Organization-wide CloudTrail
    • IAM roles instead of user keys
    • Encryption with AWS KMS keys
    • GuardDuty and Security Hub CSPM

    Operational excellence

    • Infrastructure as code
    • Runbooks for common alerts
    • Deployment rollback
    • Dashboards per service

    What changes when the workload holds health data?

    When a workload holds health data, we map each security pillar finding to the HIPAA Security Rule safeguard it affects, so one remediation backlog serves both the review and your HIPAA risk analysis.

    We add the Healthcare Industry Lens to the review and tag findings with safeguards such as access control, audit controls or transmission security.

    Running PHI environments day to day, with runbooks and monthly reporting, sits with our cloud managed services for healthcare.

    Re-review cadence

    When should a workload be reviewed again?

    A review describes the system on the day it was done. We re-review when one of these happens.

    • A new Region, account structure or landing zone change
    • A move to a new compute model, such as containers or serverless
    • A new class of data arrives, such as payment card or health data
    • A major incident or a failed recovery test
    • A new AWS lens becomes relevant to the workload
    • At least once a year for production workloads, even when nothing large has changed

    Who asks for a Well-Architected review?

    Teams preparing for growth

    A check of reliability and performance before a launch, a large customer or a traffic peak.

    Startups after their first year on AWS

    AWS Well-Architected alignment for startups is usually about closing the security and recovery gaps left by a fast first build.

    Health and financial products

    A review with the healthcare or financial services lens ahead of a customer or regulator review.

    SaaS companies onboarding larger tenants

    The SaaS Lens applied to tenant isolation, noisy neighbours and per-tenant metrics.

    Teams that inherited an AWS estate

    A clear list of what is risky, what is fine and what to fix first.

    Scope

    What drives the scope of a Well-Architected review?

    We do not publish prices. These are the factors we size on the scoping call.

    Number of workloads

    Each workload gets its own record, lenses and milestones.

    Lenses applied

    Each added lens brings its own questions and improvement plan.

    Access to engineers and accounts

    Answers checked against the accounts take longer than interviews alone, and they are more accurate.

    Remediation share

    Whether we only run the review, or also fix the high and medium risk issues afterwards.

    Health or payment data

    Mapping findings to HIPAA safeguards or PCI DSS requirements adds a column to every finding.

    Re-review schedule

    One review, or a review followed by milestone re-reviews after each batch of fixes.

    References

    Which AWS references do we review against?

    The framework, lenses and rules behind the review and the remediation backlog.

    AWS Well-Architected FrameworkAWS Well-Architected ToolHealthcare Industry LensFinancial Services Industry LensSaaS LensHIPAA Security Rule
    Facts and sources

    Facts behind Well-Architected reviews

    Definitions from the AWS Well-Architected documentation and the HIPAA Security Rule that shape how we run and record a review.

    Frequently Asked Questions

    It depends on how many workloads and lenses are in scope and how quickly the engineers who know the system are available. We agree the workload list and the remediation scope on the scoping call.