[{"data":1,"prerenderedAt":726},["ShallowReactive",2],{"blog-en-cloud-audit-and-due-diligence-checklist":3},{"_path":4,"_dir":5,"_draft":6,"_partial":6,"_locale":7,"title":8,"description":9,"date":10,"slug":11,"author":12,"image":13,"tags":14,"language":20,"format":21,"translationKey":22,"body":23,"_type":720,"_id":721,"_source":722,"_file":723,"_stem":724,"_extension":725},"/en/blog/2026-06-29-cloud-audit-and-due-diligence-checklist","blog",false,"","Before an audit or due diligence: a cloud checklist to avoid improvising","What to review in AWS before an audit, funding round, enterprise sale or technical due diligence, without turning it into promises of compliance.","2026-06-29","cloud-audit-and-due-diligence-checklist","Alejandro Rodríguez","/og-image.png",[15,16,17,18,19],"aws","audit","security","due-diligence","reliability","en","checklist","cloud-audit-due-diligence-checklist",{"type":24,"children":25,"toc":706},"root",[26,34,40,45,50,57,62,67,72,112,117,123,128,133,138,171,176,182,187,192,225,230,236,241,246,284,289,295,300,305,310,343,348,354,359,364,397,402,408,413,418,456,461,467,472,477,515,520,526,531,536,581,586,592,597,602,607,613],{"type":27,"tag":28,"props":29,"children":31},"element","h1",{"id":30},"before-an-audit-or-due-diligence-a-cloud-checklist-to-avoid-improvising",[32],{"type":33,"value":8},"text",{"type":27,"tag":35,"props":36,"children":37},"p",{},[38],{"type":33,"value":39},"An audit, a funding round or an enterprise sale can become complicated for a basic reason: nobody can confidently explain who has access to what, which logs exist, which backups have been tested, or what would happen if a critical component stopped responding.",{"type":27,"tag":35,"props":41,"children":42},{},[43],{"type":33,"value":44},"If your team runs production on AWS, preparing does not mean making the account look good or promising compliance. It means arriving with an honest assessment of the technical state, reasonable evidence and a prioritized list of risks, decisions and next steps.",{"type":27,"tag":35,"props":46,"children":47},{},[48],{"type":33,"value":49},"This checklist is not a substitute for a formal audit, a legal review or a certification process. It is a practical cloud readiness guide: what to review before an auditor, investor, enterprise customer or external technical team starts asking questions.",{"type":27,"tag":51,"props":52,"children":54},"h2",{"id":53},"_1-iam-and-access-who-can-do-what-why-and-since-when",[55],{"type":33,"value":56},"1. IAM and access: who can do what, why, and since when",{"type":27,"tag":35,"props":58,"children":59},{},[60],{"type":33,"value":61},"The first question is uncomfortably simple: who can access the account, and what can they do?",{"type":27,"tag":35,"props":63,"children":64},{},[65],{"type":33,"value":66},"In AWS, IAM accumulates history. Old users, roles created for a temporary integration, broad permissions nobody dares to touch, active credentials for projects that no longer exist, and third-party access without a clear owner.",{"type":27,"tag":35,"props":68,"children":69},{},[70],{"type":33,"value":71},"Before technical due diligence, it is not enough to say that IAM exists. You should be able to show or explain:",{"type":27,"tag":73,"props":74,"children":75},"ul",{},[76,82,87,92,97,102,107],{"type":27,"tag":77,"props":78,"children":79},"li",{},[80],{"type":33,"value":81},"how people access the environment;",{"type":27,"tag":77,"props":83,"children":84},{},[85],{"type":33,"value":86},"whether the root user is protected and is not used for day-to-day operations;",{"type":27,"tag":77,"props":88,"children":89},{},[90],{"type":33,"value":91},"which users, roles or credentials remain active without a clear need;",{"type":27,"tag":77,"props":93,"children":94},{},[95],{"type":33,"value":96},"where permissions are too broad;",{"type":27,"tag":77,"props":98,"children":99},{},[100],{"type":33,"value":101},"what access vendors, freelancers or external tools have;",{"type":27,"tag":77,"props":103,"children":104},{},[105],{"type":33,"value":106},"whether MFA and temporary credentials are used where appropriate;",{"type":27,"tag":77,"props":108,"children":109},{},[110],{"type":33,"value":111},"who approves new permissions and who reviews existing ones.",{"type":27,"tag":35,"props":113,"children":114},{},[115],{"type":33,"value":116},"The useful question is not whether something is secure in the abstract. It is whether you can justify why each relevant access exists and what impact it would have if misused.",{"type":27,"tag":51,"props":118,"children":120},{"id":119},"_2-logging-and-traceability-being-able-to-reconstruct-important-changes",[121],{"type":33,"value":122},"2. Logging and traceability: being able to reconstruct important changes",{"type":27,"tag":35,"props":124,"children":125},{},[126],{"type":33,"value":127},"An external review raises a traceability question: can you find out who changed this, when, and from where?",{"type":27,"tag":35,"props":129,"children":130},{},[131],{"type":33,"value":132},"CloudTrail can help record the activity of users, roles and services in AWS. AWS Config can provide configuration history and relationships between the resource types for which recording has been enabled. CloudWatch Logs and other logging systems can help you understand application and infrastructure behavior.",{"type":27,"tag":35,"props":134,"children":135},{},[136],{"type":33,"value":137},"The important qualification: having those names in the account does not mean you have sufficient evidence. Review:",{"type":27,"tag":73,"props":139,"children":140},{},[141,146,151,156,161,166],{"type":27,"tag":77,"props":142,"children":143},{},[144],{"type":33,"value":145},"whether CloudTrail is configured with the scope and retention you need;",{"type":27,"tag":77,"props":147,"children":148},{},[149],{"type":33,"value":150},"which events are covered and which are not;",{"type":27,"tag":77,"props":152,"children":153},{},[154],{"type":33,"value":155},"whether critical logs are centralized or scattered;",{"type":27,"tag":77,"props":157,"children":158},{},[159],{"type":33,"value":160},"who can read, delete or modify evidence;",{"type":27,"tag":77,"props":162,"children":163},{},[164],{"type":33,"value":165},"how long relevant data is retained;",{"type":27,"tag":77,"props":167,"children":168},{},[169],{"type":33,"value":170},"whether there is a reasonable way to investigate changes in production.",{"type":27,"tag":35,"props":172,"children":173},{},[174],{"type":33,"value":175},"The goal is not to keep infinite logs just in case. It is to avoid depending on tribal knowledge or scattered screenshots when a serious question arises. CloudTrail event history shows the last 90 days of management events per region; it does not amount to coverage of all events or a long-term archive. Check trails or event data stores and their event selection when the scope requires more. Do not reduce retention or delete logs without validating retention requirements and ongoing investigations.",{"type":27,"tag":51,"props":177,"children":179},{"id":178},"_3-public-exposure-knowing-what-is-exposed-and-why",[180],{"type":33,"value":181},"3. Public exposure: knowing what is exposed and why",{"type":27,"tag":35,"props":183,"children":184},{},[185],{"type":33,"value":186},"Network exposure accumulates exceptions: a security group opened for a test, an S3 bucket that was supposed to be temporarily public, an abandoned endpoint, a load balancer without a clear owner, or a legacy rule nobody knows whether they can close.",{"type":27,"tag":35,"props":188,"children":189},{},[190],{"type":33,"value":191},"Before an audit or due diligence, review at least:",{"type":27,"tag":73,"props":193,"children":194},{},[195,200,205,210,215,220],{"type":27,"tag":77,"props":196,"children":197},{},[198],{"type":33,"value":199},"security groups with broad address ranges, especially on administrative ports;",{"type":27,"tag":77,"props":201,"children":202},{},[203],{"type":33,"value":204},"resources accessible from the Internet and the justification for that access;",{"type":27,"tag":77,"props":206,"children":207},{},[208],{"type":33,"value":209},"S3 buckets and their Block Public Access configuration;",{"type":27,"tag":77,"props":211,"children":212},{},[213],{"type":33,"value":214},"load balancers, public endpoints and legacy rules;",{"type":27,"tag":77,"props":216,"children":217},{},[218],{"type":33,"value":219},"documented exceptions: what is open, to whom, and why;",{"type":27,"tag":77,"props":221,"children":222},{},[223],{"type":33,"value":224},"external dependencies connecting into or out of the platform.",{"type":27,"tag":35,"props":226,"children":227},{},[228],{"type":33,"value":229},"Not every public resource is a mistake. A web application will have exposed components. What is worrying is not knowing what is publicly accessible, who decided it should be, and which controls compensate for that design.",{"type":27,"tag":51,"props":231,"children":233},{"id":232},"_4-backups-checking-recovery-not-just-existence",[234],{"type":33,"value":235},"4. Backups: checking recovery, not just existence",{"type":27,"tag":35,"props":237,"children":238},{},[239],{"type":33,"value":240},"A snapshot, a backup job or an AWS Backup policy can look reassuring. But a backup that has never been restored is still a hypothesis.",{"type":27,"tag":35,"props":242,"children":243},{},[244],{"type":33,"value":245},"The review should answer specific questions:",{"type":27,"tag":73,"props":247,"children":248},{},[249,254,259,264,269,274,279],{"type":27,"tag":77,"props":250,"children":251},{},[252],{"type":33,"value":253},"which critical data and services are backed up;",{"type":27,"tag":77,"props":255,"children":256},{},[257],{"type":33,"value":258},"how often backups are made;",{"type":27,"tag":77,"props":260,"children":261},{},[262],{"type":33,"value":263},"how long they are retained;",{"type":27,"tag":77,"props":265,"children":266},{},[267],{"type":33,"value":268},"who receives alerts if a backup fails;",{"type":27,"tag":77,"props":270,"children":271},{},[272],{"type":33,"value":273},"when the last tested restore took place;",{"type":27,"tag":77,"props":275,"children":276},{},[277],{"type":33,"value":278},"which RPO (tolerable data loss measured in time) and RTO (target recovery time) the business needs, and which results have been measured in tests;",{"type":27,"tag":77,"props":280,"children":281},{},[282],{"type":33,"value":283},"which dependencies are not covered by the current plan.",{"type":27,"tag":35,"props":285,"children":286},{},[287],{"type":33,"value":288},"You do not need to turn the backup plan into a perfect disaster recovery program overnight. But you should distinguish three states: what is backed up and tested, what is backed up but not tested, and what simply has no clear strategy.",{"type":27,"tag":51,"props":290,"children":292},{"id":291},"_5-continuity-and-dr-do-not-confuse-high-availability-with-recovery",[293],{"type":33,"value":294},"5. Continuity and DR: do not confuse high availability with recovery",{"type":27,"tag":35,"props":296,"children":297},{},[298],{"type":33,"value":299},"High availability, backups and disaster recovery are not the same thing.",{"type":27,"tag":35,"props":301,"children":302},{},[303],{"type":33,"value":304},"Having several containers behind a load balancer can help with certain failures. Having backups can help with data loss or corruption. Having a disaster recovery plan means knowing how to recover the service from more serious scenarios and what degradation the business accepts.",{"type":27,"tag":35,"props":306,"children":307},{},[308],{"type":33,"value":309},"A realistic review should document scenarios such as:",{"type":27,"tag":73,"props":311,"children":312},{},[313,318,323,328,333,338],{"type":27,"tag":77,"props":314,"children":315},{},[316],{"type":33,"value":317},"failure of an instance, node, task or managed service;",{"type":27,"tag":77,"props":319,"children":320},{},[321],{"type":33,"value":322},"data loss or corruption;",{"type":27,"tag":77,"props":324,"children":325},{},[326],{"type":33,"value":327},"failure of an availability zone;",{"type":27,"tag":77,"props":329,"children":330},{},[331],{"type":33,"value":332},"unavailability of a critical dependency;",{"type":27,"tag":77,"props":334,"children":335},{},[336],{"type":33,"value":337},"human error during an infrastructure change;",{"type":27,"tag":77,"props":339,"children":340},{},[341],{"type":33,"value":342},"the need to rebuild or restore an environment from scratch.",{"type":27,"tag":35,"props":344,"children":345},{},[346],{"type":33,"value":347},"The answer does not have to be multi-region for everything. For many teams, that would be expensive, unnecessary or difficult to operate well. The useful answer is to explain which architecture exists, which scenarios it covers, which it does not, and which risks are consciously accepted.",{"type":27,"tag":51,"props":349,"children":351},{"id":350},"_6-observability-and-alerts-signals-someone-can-use",[352],{"type":33,"value":353},"6. Observability and alerts: signals someone can use",{"type":27,"tag":35,"props":355,"children":356},{},[357],{"type":33,"value":358},"To prepare for technical due diligence, you should assess observability beyond simply having dashboards: check whether the team can detect, understand and respond to relevant problems. The specific scope will depend on the external review.",{"type":27,"tag":35,"props":360,"children":361},{},[362],{"type":33,"value":363},"Review whether a minimum set of signals exists for critical services:",{"type":27,"tag":73,"props":365,"children":366},{},[367,372,377,382,387,392],{"type":27,"tag":77,"props":368,"children":369},{},[370],{"type":33,"value":371},"infrastructure and application metrics;",{"type":27,"tag":77,"props":373,"children":374},{},[375],{"type":33,"value":376},"logs that can be queried during incidents;",{"type":27,"tag":77,"props":378,"children":379},{},[380],{"type":33,"value":381},"actionable alarms, not constant noise;",{"type":27,"tag":77,"props":383,"children":384},{},[385],{"type":33,"value":386},"a clear owner for each important alert;",{"type":27,"tag":77,"props":388,"children":389},{},[390],{"type":33,"value":391},"dashboards that explain service health, not just CPU;",{"type":27,"tag":77,"props":393,"children":394},{},[395],{"type":33,"value":396},"the relationship between deployments, changes and incidents.",{"type":27,"tag":35,"props":398,"children":399},{},[400],{"type":33,"value":401},"CloudWatch can cover part of this within AWS, but the particular tool matters less than operational discipline. An alert nobody understands is not strong evidence. Neither is a dashboard nobody looks at.",{"type":27,"tag":51,"props":403,"children":405},{"id":404},"_7-infrastructure-as-code-and-changes-knowing-the-source-of-truth",[406],{"type":33,"value":407},"7. Infrastructure as code and changes: knowing the source of truth",{"type":27,"tag":35,"props":409,"children":410},{},[411],{"type":33,"value":412},"Terraform, or any infrastructure-as-code approach, helps when it allows you to review changes, understand dependencies and reduce invisible manual configuration. But it should not be sold as a religion.",{"type":27,"tag":35,"props":414,"children":415},{},[416],{"type":33,"value":417},"Before an audit or external review, what matters is knowing:",{"type":27,"tag":73,"props":419,"children":420},{},[421,426,431,436,441,446,451],{"type":27,"tag":77,"props":422,"children":423},{},[424],{"type":33,"value":425},"which parts of the infrastructure are in code;",{"type":27,"tag":77,"props":427,"children":428},{},[429],{"type":33,"value":430},"which parts are changed manually;",{"type":27,"tag":77,"props":432,"children":433},{},[434],{"type":33,"value":435},"how changes are reviewed before touching production;",{"type":27,"tag":77,"props":437,"children":438},{},[439],{"type":33,"value":440},"whether there are pull requests, plans, approvals or an equivalent workflow;",{"type":27,"tag":77,"props":442,"children":443},{},[444],{"type":33,"value":445},"where drift may exist between Terraform and reality;",{"type":27,"tag":77,"props":447,"children":448},{},[449],{"type":33,"value":450},"who maintains modules, state files and secrets;",{"type":27,"tag":77,"props":452,"children":453},{},[454],{"type":33,"value":455},"which recent changes could explain incidents or risks.",{"type":27,"tag":35,"props":457,"children":458},{},[459],{"type":33,"value":460},"If everything is in Terraform but nobody dares to run a plan, there is not as much maturity as it seems. If part of the platform is manual but documented, controlled and has clear owners, the risk may be manageable while gradual improvements are prioritized.",{"type":27,"tag":51,"props":462,"children":464},{"id":463},"_8-documentation-and-ownership-answering-without-improvising",[465],{"type":33,"value":466},"8. Documentation and ownership: answering without improvising",{"type":27,"tag":35,"props":468,"children":469},{},[470],{"type":33,"value":471},"A technical review can stall because of missing explanations, even when controls exist. The platform may be reasonably sound, but if nobody can describe it, it looks more fragile than it is.",{"type":27,"tag":35,"props":473,"children":474},{},[475],{"type":33,"value":476},"Before external pressure arrives, prepare the basics:",{"type":27,"tag":73,"props":478,"children":479},{},[480,485,490,495,500,505,510],{"type":27,"tag":77,"props":481,"children":482},{},[483],{"type":33,"value":484},"a current architecture diagram;",{"type":27,"tag":77,"props":486,"children":487},{},[488],{"type":33,"value":489},"a list of critical services and dependencies;",{"type":27,"tag":77,"props":491,"children":492},{},[493],{"type":33,"value":494},"owners for each area: infrastructure, security, data, deployments and incidents;",{"type":27,"tag":77,"props":496,"children":497},{},[498],{"type":33,"value":499},"runbooks for sensitive operations;",{"type":27,"tag":77,"props":501,"children":502},{},[503],{"type":33,"value":504},"important architectural decisions and their context;",{"type":27,"tag":77,"props":506,"children":507},{},[508],{"type":33,"value":509},"known risks and the status of their mitigation;",{"type":27,"tag":77,"props":511,"children":512},{},[513],{"type":33,"value":514},"an improvement roadmap with realistic priorities.",{"type":27,"tag":35,"props":516,"children":517},{},[518],{"type":33,"value":519},"Useful documentation does not have to be long. It has to prevent every answer from depending on the person who remembers how everything was put together.",{"type":27,"tag":51,"props":521,"children":523},{"id":522},"_9-prioritization-not-everything-gets-fixed-before-the-audit",[524],{"type":33,"value":525},"9. Prioritization: not everything gets fixed before the audit",{"type":27,"tag":35,"props":527,"children":528},{},[529],{"type":33,"value":530},"A common mistake is to turn a preliminary review into a race to fix everything. That tends to produce rushed changes, little evidence and more operational risk.",{"type":27,"tag":35,"props":532,"children":533},{},[534],{"type":33,"value":535},"A more useful outcome is to classify findings into four groups:",{"type":27,"tag":537,"props":538,"children":539},"ol",{},[540,551,561,571],{"type":27,"tag":77,"props":541,"children":542},{},[543,549],{"type":27,"tag":544,"props":545,"children":546},"strong",{},[547],{"type":33,"value":548},"Actual blockers",{"type":33,"value":550},": risks that should be remediated before the external review.",{"type":27,"tag":77,"props":552,"children":553},{},[554,559],{"type":27,"tag":544,"props":555,"children":556},{},[557],{"type":33,"value":558},"Significant accepted risks",{"type":33,"value":560},": decisions that are not being corrected yet but are explained with context.",{"type":27,"tag":77,"props":562,"children":563},{},[564,569],{"type":27,"tag":544,"props":565,"children":566},{},[567],{"type":33,"value":568},"Quick improvements",{"type":33,"value":570},": low-risk adjustments that reduce uncertainty.",{"type":27,"tag":77,"props":572,"children":573},{},[574,579],{"type":27,"tag":544,"props":575,"children":576},{},[577],{"type":33,"value":578},"Follow-up roadmap",{"type":33,"value":580},": important work that requires design, testing or budget.",{"type":27,"tag":35,"props":582,"children":583},{},[584],{"type":33,"value":585},"This changes the tone of the conversation. Instead of promising that everything is perfect, the team can say: this is covered, this is in progress, this is accepted for now, and this requires a business decision.",{"type":27,"tag":51,"props":587,"children":589},{"id":588},"arriving-prepared-does-not-mean-promising-compliance",[590],{"type":33,"value":591},"Arriving prepared does not mean promising compliance",{"type":27,"tag":35,"props":593,"children":594},{},[595],{"type":33,"value":596},"Serious preparation before an audit or due diligence is not about manufacturing a flawless image of AWS. It is about understanding the platform, showing reasonable evidence and speaking clearly about risks.",{"type":27,"tag":35,"props":598,"children":599},{},[600],{"type":33,"value":601},"IAM, logging, public exposure, backups, recovery, observability, IaC, documentation and ownership are not a guarantee of compliance. They are areas where a platform shows whether it is operated thoughtfully or depends too much on memory, urgency and luck.",{"type":27,"tag":35,"props":603,"children":604},{},[605],{"type":33,"value":606},"If your team is preparing for an audit, a funding round or an enterprise sale and needs to separate real risks from operational noise, AstralDeploy can help with an AWS architecture and security review focused on evidence, priorities and next steps.",{"type":27,"tag":51,"props":608,"children":610},{"id":609},"sources-consulted",[611],{"type":33,"value":612},"Sources consulted",{"type":27,"tag":73,"props":614,"children":615},{},[616,629,640,651,662,673,684,695],{"type":27,"tag":77,"props":617,"children":618},{},[619,621],{"type":33,"value":620},"AWS IAM — Security best practices, root, MFA, temporary credentials and least privilege: ",{"type":27,"tag":622,"props":623,"children":627},"a",{"href":624,"rel":625},"https://docs.aws.amazon.com/IAM/latest/UserGuide/best-practices.html",[626],"nofollow",[628],{"type":33,"value":624},{"type":27,"tag":77,"props":630,"children":631},{},[632,634],{"type":33,"value":633},"AWS CloudTrail — Scope of event history and logging options: ",{"type":27,"tag":622,"props":635,"children":638},{"href":636,"rel":637},"https://docs.aws.amazon.com/awscloudtrail/latest/userguide/cloudtrail-user-guide.html",[626],[639],{"type":33,"value":636},{"type":27,"tag":77,"props":641,"children":642},{},[643,645],{"type":33,"value":644},"AWS Config — Configuration history and recorded resource types: ",{"type":27,"tag":622,"props":646,"children":649},{"href":647,"rel":648},"https://docs.aws.amazon.com/config/latest/developerguide/WhatIsConfig.html",[626],[650],{"type":33,"value":647},{"type":27,"tag":77,"props":652,"children":653},{},[654,656],{"type":33,"value":655},"Amazon S3 — Block Public Access: ",{"type":27,"tag":622,"props":657,"children":660},{"href":658,"rel":659},"https://docs.aws.amazon.com/AmazonS3/latest/userguide/access-control-block-public-access.html",[626],[661],{"type":33,"value":658},{"type":27,"tag":77,"props":663,"children":664},{},[665,667],{"type":33,"value":666},"AWS Well-Architected — REL 9, backups and recovery testing: ",{"type":27,"tag":622,"props":668,"children":671},{"href":669,"rel":670},"https://docs.aws.amazon.com/wellarchitected/latest/framework/rel-09.html",[626],[672],{"type":33,"value":669},{"type":27,"tag":77,"props":674,"children":675},{},[676,678],{"type":33,"value":677},"AWS Well-Architected — REL 13, RPO, RTO, disaster recovery strategies and testing: ",{"type":27,"tag":622,"props":679,"children":682},{"href":680,"rel":681},"https://docs.aws.amazon.com/wellarchitected/latest/framework/rel-13.html",[626],[683],{"type":33,"value":680},{"type":27,"tag":77,"props":685,"children":686},{},[687,689],{"type":33,"value":688},"AWS Well-Architected — Monitoring, observability and operational responsibilities: ",{"type":27,"tag":622,"props":690,"children":693},{"href":691,"rel":692},"https://docs.aws.amazon.com/wellarchitected/latest/management-and-governance-guide/observability.html",[626],[694],{"type":33,"value":691},{"type":27,"tag":77,"props":696,"children":697},{},[698,700],{"type":33,"value":699},"HashiCorp Terraform — Planning and reviewing changes: ",{"type":27,"tag":622,"props":701,"children":704},{"href":702,"rel":703},"https://developer.hashicorp.com/terraform/cli/commands/plan",[626],[705],{"type":33,"value":702},{"title":7,"searchDepth":707,"depth":707,"links":708},2,[709,710,711,712,713,714,715,716,717,718,719],{"id":53,"depth":707,"text":56},{"id":119,"depth":707,"text":122},{"id":178,"depth":707,"text":181},{"id":232,"depth":707,"text":235},{"id":291,"depth":707,"text":294},{"id":350,"depth":707,"text":353},{"id":404,"depth":707,"text":407},{"id":463,"depth":707,"text":466},{"id":522,"depth":707,"text":525},{"id":588,"depth":707,"text":591},{"id":609,"depth":707,"text":612},"markdown","content:en:blog:2026-06-29-cloud-audit-and-due-diligence-checklist.md","content","en/blog/2026-06-29-cloud-audit-and-due-diligence-checklist.md","en/blog/2026-06-29-cloud-audit-and-due-diligence-checklist","md",1795108324686]