Azure subscription sprawl is governance debt
How Azure tenants sprawl one decision at a time, a management-group layout a small team can run, and the order to clean up RBAC without breaking anyone.
azure · governance
Field notes
What moves delivery speed, reliability, and cloud spend, and what just adds ceremony. Written when we have something to say.
Why your AWS bill keeps growing even when traffic doesn't
The specific mechanisms that make an AWS bill climb while traffic stays flat, and why per-team cost visibility changes behavior more than any dashboard.
aws · cost-optimization
Do you need Kubernetes?
What a Kubernetes cluster really costs to operate, what small teams should run instead, and a four-question heuristic for when it earns its keep.
kubernetes · platform-engineering
Terraform isn't the hard part. Managing infrastructure changes is.
Why teams that write HCL fluently still fear terraform apply: plan review culture, state drift, blast radius, CI-run applies, and what not to codify yet.
terraform · infrastructure-as-code
Platform engineering: when (and when not) to build an internal developer platform
IDPs are the most oversold idea in DevOps right now. Here's the test we use to decide whether a team needs one, and the cheaper moves that usually come first.
platform-engineering · idp
What we look for during an infrastructure review
The review lens we bring to a client account: what read-only access shows on day one, the five categories we walk, and what a real finding looks like.
infrastructure-review · cloud-infrastructure
Your CI pipeline is slower than it needs to be
Caching, test sharding, image layering, runner sizing: the four levers that reliably shrink CI feedback loops, in the order worth pulling them.
ci-cd · developer-experience
SLOs your engineers will use: starting SRE without the ceremony
Most SLO rollouts die in a spreadsheet. Start with four user journeys, alerts wired to burn rate, and one review ritual. Skip the rest until it hurts.
sre · reliability