Chaos Engineering on AWS: Fault Injection Simulator Guide
You don’t know your system is resilient until you’ve broken it on purpose. I believed our payment processing service was fault …
Read article →
Since 2007 · 1782 articles · 362 topics
Cloud Architecture, DevOps & Systems Engineering
Hands-on technical guides on AWS, Kubernetes, Terraform, Rust, Python and Go — written from production, not theory. By Andrew Odendaal.
This is a working engineer's notebook, not a content farm. Every article came out of a real production problem — something that broke at 2am, a migration that wouldn't behave, a tool that didn't do what its docs claimed. I've been publishing here since 2007.
You'll find everything from current AWS architecture deep-dives back through old PHP and Flash war stories. The technology dates, the lessons rarely do. If you're searching for a specific error, a comparison between two tools, or a deeper take on something the official docs gloss over — that's what this site is for.
Browse by topic
The clusters I keep returning to. Each card links to a topic hub plus the three articles I'd point a colleague at first.
VPC design, ECS & EKS, Lambda internals, Step Functions, multi-account strategy, cost optimization. Production-grade patterns, not certification rehashes.
- 15 AWS cost optimization techniques that actually work
- Lambda cold starts: causes, measurement, mitigation
- ECS vs EKS: choosing the right container orchestrator
RBAC, network policies, operators, autoscaling, multi-cluster, eBPF, Gateway API. Hard-won notes from running production clusters.
- RBAC deep dive: securing multi-tenant clusters
- Network policies: a practical security guide
- eBPF: next-generation observability and security
State management, module design, testing strategies, drift detection. Plus the CDK vs Terraform tradeoffs nobody talks about honestly.
- State management best practices
- Module design patterns for reusable infrastructure
- AWS CDK vs Terraform: a practical comparison
Error handling, async runtimes, Tokio internals, WebAssembly. Why Rust matters for cloud engineers, with real examples.
- Error handling patterns for production applications
- Async runtime deep dive: Tokio architecture
- WebAssembly: high-performance web applications
asyncio in production, type hints at scale, performance profiling, modern packaging with uv and Poetry, data pipelines.
- Async programming: asyncio, tasks, real-world patterns
- Type hints and static analysis in production
- Data pipelines with Polars and DuckDB
GitOps with ArgoCD, distributed tracing, SLOs, chaos engineering, container security, platform engineering. Modern operational practice.
- GitOps with ArgoCD: from zero to production
- Distributed tracing with OpenTelemetry
- Implementing SLOs and error budgets in practice
Latest articles
Python Data Pipelines with Polars and DuckDB
I’m going to say something that’ll upset a lot of people: pandas had its run. Polars is just better. I don’t mean that …
Read article →Why I Built Wyn: A Programming Language That Compiles to C
I’ve spent years writing Python for DevOps tooling and Go for services. Python is a joy to write but painfully slow for anything …
Read article →Kubernetes Gateway API: The Future of Ingress
Gateway API is what Ingress should have been from day one. I don’t say that lightly. I’ve spent years wrangling Kubernetes …
Read article →AWS Bedrock: Building AI Applications with Foundation Models
Bedrock is AWS finally getting AI right. I don’t say that lightly. I’ve watched AWS stumble through SageMaker’s …
Read article →Infrastructure Drift Detection and Remediation
If you’re not running scheduled terraform plan, you have drift. You just don’t know it yet. I learned this the hard way. A …
Read article →What you'll find here
The articles are deliberately specific. A post on Kubernetes RBAC walks through the actual permission boundaries you'd hit in a multi-tenant cluster, with the YAML and the failure modes. A post on Terraform state explains the locking strategy that survived three production incidents. A post on Rust error handling shows the pattern after rewriting the same module four times.
I avoid two things: theoretical purity and listicle filler. If a recommendation sounds good but I haven't shipped it, it doesn't go in. If something is genuinely a bad idea — even if it's trendy — I'll say so. The hope is you find the same kind of writing useful that I find useful when I'm searching for someone else's hard-won notes.
All topics
Twenty most-written-about. The full tag index has every topic from 362 categories.