Posts on DevOps and developer tooling — Docker, Git, CI/CD, deployment, and the dozens of small workflow choices that compound into a great developer experience.
How to deploy without downtime in 2026. Blue/green, canary, rolling, feature flags, db migrations the right way, and the patterns that survive Friday afternoons.
Practical incident response in 2026. Severity levels, IC role, comms cadence, runbooks, blameless postmortems, action item tracking, and the cultural shifts that produce real learning.
Where Kubernetes stands in 2026: eBPF-default networking, Gateway API, sidecarless meshes, KEDA, Argo, and the operational realities you’ll face.
Picking a reproducible dev-environment tool in 2026. Nix flakes (deepest), Devbox (Nix made friendly), mise (lightweight version manager), devcontainers (Docker-based), and which fits which team.
Picking an IaC tool in 2026. Pulumi (real programming languages), Terraform (HashiCorp’s BSL-licensed original), OpenTofu (open-source fork). Tradeoffs, ecosystem, state, and how to migrate.
Why eBPF won the cloud-native networking layer in 2026. Cilium architecture, sidecarless service mesh, network policy at L7, mTLS, observability with Hubble + Tetragon + Pixie, and the migration path from Istio.
How OpenTelemetry actually works in 2026. The collector, traces / metrics / logs / context propagation, instrumenting Python, Go, and Rust services, sampling that doesn’t kill recall, and the production-grade end-to-end setup.
A grounded look at WebAssembly on Kubernetes in 2026. WASI 0.2’s component model, Spin and wasmCloud, runwasi shims, the workloads where WASM is genuinely better than containers, and the ones where it isn’t.
How modern supply chain security actually works — SBOMs, SLSA levels, signing with Sigstore/cosign, attestations, and a practical CI pipeline that protects against typosquatting, dependency hijacks, and tampered builds.
A short, practical guide to SLOs and error budgets for application developers. Choose the right SLI, pick targets you can actually defend, calculate the budget, and use it to drive feature-velocity vs. reliability tradeoffs.