Archive
32 posts, newest first. Nothing is removed; corrections are dated in place.
2026
-
Cost governance that engineers acceptCost 12 min
Cloud cost is an architectural property, not a monthly report. Tags that hold, priced switches, endpoint arithmetic and a cost comment on every pull request.
-
Data platform governance with Lake FormationPlatform 9 min
A three-zone lake, LF-Tags and a deny-direct-S3 policy turn data governance into something enforced at query time instead of described in a spreadsheet.
-
An EKS platform without a platform teamPlatform 11 min
A worked EKS golden path: a managed node group with Karpenter, ArgoCD, External Secrets, promotion by approval, and what Kubernetes costs you to operate.
-
The golden path for new servicesPlatform 12 min
A golden path turns a platform team from a ticket queue into a paved road: what it must include, the new-repository flow, and the 18 pitfalls to design out.
-
Terraform state at enterprise scaleDelivery 12 min
Terraform state is a production system: one key per unit, a real lock, versioning, and a restore drill. Here is what to do when state and reality diverge.
-
OpenTofu and Terragrunt for enterprise IaCDelivery 11 min
OpenTofu and Terragrunt give enterprise IaC a license you control and a dependency graph that plans an entire new environment before any of it exists.
-
Evidence automation beats screenshot seasonCompliance 9 min
Audit evidence has a shelf life, and most findings services keep only 90 days. Here is how to make infrastructure produce evidence that survives the period.
-
Infrastructure due diligence: what buyers look forCompliance 10 min
Investors, acquirers and enterprise procurement inspect the same six things. Here is what each finding signals, and a self-audit to run before they arrive.
-
Answering a 300-question security questionnaireCompliance 9 min
A security questionnaire is a revenue blocker answered by your most expensive engineers. Here is how to answer with evidence and never write the answer twice.
-
CMMC 2.0 for defense suppliersCompliance 9 min
CMMC Level 1 and Level 2 decide whether a supplier can hold the contract. Here is what infrastructure covers, what it cannot, and where it gets hard.
-
PCI DSS 4.0: scope reduction by accountCompliance 9 min
Account isolation is the cheapest PCI scope reduction available, and the most over-claimed. Here is what it removes, what stays in scope, and what proves it.
-
HIPAA on AWS: what the foundation must carryCompliance 9 min
The HIPAA Security Rule splits across your policies, your cloud provider and your infrastructure. Here is the part infrastructure owns, and how it is proven.
-
Zero trust access for engineersSecurity 9 min
How people should reach production: federated sessions instead of users, no long-lived keys anywhere, and access that ends the day someone leaves.
-
Built for 3 AM: the operability propertiesOperations 7 min
Operability is what a system needs before it has an incident. Here are the properties that decide whether the second-best engineer can run it alone.
-
Secrets management without a vault clusterSecurity 10 min
Encrypted secrets in Git, one key per stage, and a sync to parameter storage: why repository write access still does not let anyone read production.
-
Tamper-resistant audit logging, honestlyCompliance 10 min
A separate account, an organization-wide trail with file validation, and a policy that forbids stopping it — plus what tamper-resistant does not mean.
-
Nine security services, one accountSecurity 9 min
What GuardDuty, Security Hub, Config, CloudTrail and their siblings each actually catch, what they cost, and the order to switch them on.
-
Production readiness is a velocity multiplierCost 6 min
Production engineering is usually sold as risk reduction. The stronger argument is speed: teams without a foundation spend their capacity on rework.
-
IP address planning that survives growthArchitecture 9 min
Overlapping CIDRs are the cheapest mistake to make and one of the most expensive to fix. Here is how to plan an address space once and enforce it in CI.
-
Hub and spoke networking with Transit GatewayArchitecture 10 min
Spoke VPCs with no NAT, one centralized egress path, and blackhole routes between production and non-production: what a hub and spoke topology buys.
-
Permission sets that survive an auditSecurity 9 min
Least privilege fails when people cannot do their jobs. Here is an identity model with read-only production and an access review that produces evidence.
-
Service control policies that holdSecurity 9 min
A guardrail works only if engineers cannot route around it. Here is how to design, exempt and test a service control policy before it blocks production.
-
Drift is a fact, not a failurePlatform 7 min
Infrastructure drift is not a sign of a broken team. It is information about what your environment needs. Detect it, triage it, and never auto-revert it.
-
Control Tower vs a code-owned landing zoneArchitecture 9 min
AWS Control Tower gives you the account and guardrail layer at no license fee. Here is where a code-owned foundation wins, and where it honestly does not.
-
Multi-account strategy for enterprisesArchitecture 9 min
The account boundary is the first security control you get. Here is how to decide what deserves its own account, and how the structure survives growth.
-
If deployment requires a hero, it isn't automatedDelivery 6 min
A deploy that needs the right person, the right laptop or the right afternoon is a manual process with a script attached. Here is the difference.
-
What production-ready actually meansArchitecture 8 min
Production-ready is not a feeling. It is eight properties, each with a question you can answer about your own estate this week, with evidence.
-
The true cost of building your own landing zoneArchitecture 10 min
A cloud foundation built in-house costs far more than the estimate says. Here is the work nobody scopes, and when building it yourself is still right.
-
SOC 2: what your infrastructure has to produceCompliance 7 min
An audit is an evidence problem. Here are the evidence sources cloud infrastructure should generate on its own, and why readiness is not certification.
-
Zero long-lived cloud credentials: OIDC in CI/CDSecurity 6 min
Static cloud keys in CI are a liability you chose to keep. OIDC federation replaces them with short-lived tokens your pipeline requests per job.
-
The multi-account landing zone, explainedArchitecture 7 min
Account boundaries are the first security control you get and the hardest to add later. Here is what each account in a landing zone is for, and why.
-
Backups you haven't restored are assumptionsReliability 8 min
A backup nobody has restored is a hypothesis, not a control. Here is how to run a restore drill and set an RTO and an RPO you can defend.
By subject area
By tag
Scheduled
Written, reviewed and merged. Each goes live on the first build after its date — the site rebuilds on a schedule, so no one has to press anything.
- Disaster recovery you can prove
- Day 2 operations: who owns production
- From ClickOps to infrastructure as code
- Scaling to a new region without a redesign
- What handover should look like
- Choosing a cloud foundation partner
- The GuardDuty findings that deserve a page
- A KMS key strategy for multi-account AWS
- AWS WAF rules that do not break your app
- Securing the CI/CD supply chain
- VPC flow logs: what to do with them
- Incident response for cloud infrastructure
- FedRAMP and NIST 800-53: what actually maps
- Data residency is an architecture problem
- Twelve weeks to an audit, week by week
- The healthtech path to enterprise readiness
- Fintech infrastructure under two frameworks
- Surviving a vendor security review
- Do you actually need Kubernetes?
- ECS versus EKS for enterprise workloads
- ArgoCD patterns that scale
- Karpenter versus Cluster Autoscaler
- Pod security in practice
- Multi-tenancy in a shared cluster
- A data lake an auditor accepts
- Glue, EMR or Lambda for ETL
- Data classification that holds
- AI workloads on a governed foundation
- GPU capacity without a blank check
- The analytics account pattern
- SLOs for infrastructure teams
- Multi-AZ is not a disaster recovery plan
- Chaos testing your landing zone
- Observability without a six-figure bill
- The four hundred alerts problem
- Capacity planning for accounts and regions
- Pull request workflows for infrastructure
- Testing infrastructure code without theater
- Module design that survives upgrades
- Environment parity that is worth paying for
- CI secrets without long-lived keys
- Provider and module upgrades without fear
- The CTO case for buying your foundation
- Platform team sizing and structure
- Your first 90 days as head of platform
- Build versus buy after a funding round
- The hidden cost of cheap contractors
- Infrastructure decisions that affect valuation
- When to hire your first platform engineer
- Migrating off a platform as a service
- From one AWS account to many
- Onboarding engineers to a new platform
- Runbooks people actually use
- Change management without a change board
- Access reviews that are not theater
- When your infrastructure lead quits
Find out what production would take.
A 30 minutes call with an engineer who scopes engagements, not a salesperson. Tell us what you run and what is coming: a funding round, an audit, a first enterprise customer, a migration. You leave with a next step, a fixed fee, or an honest no.