Field notes
Senior engineering perspectives from production systems — not generic cloud commentary.
Start here
Six foundational notes if you’re new to how we think.
- 1Kubernetes cost optimisation: a utilisation problem, not a price problemUtilisation first, price second
- 2DORA's four keys: a guardrail for the AI era, not a leaderboardHow we measure delivery
- 3Observability at scale: when telemetry becomes a deletion problemWhen telemetry becomes the problem
- 4FinOps for DevOps teams: cost accountability without the blameCost accountability without blame
- 5Disaster recovery in the cloud: RPO, RTO and tested restoresRestores you have actually tested
- 6AWS to GCP: what actually changes (and when not to move)What a cross-cloud move changes
Latest from production
DORA's four keys: a guardrail for the AI era, not a leaderboard
The four keys are a guardrail between speed and stability — not a score to chase.
20 Jun 2026 · 16 min read
AI is an amplifier, not a shortcut: reading the 2025 DORA signal
AI makes good delivery systems faster and weak ones weaker — the platform decides which.
16 Jun 2026 · 18 min read
Cell-based architecture: buying down blast radius
Stop partial failures from spreading: blast-radius boundaries you can drain.
12 Jun 2026 · 20 min read
Cloud migration ROI: why optimisation debt compounds — and how to break the cycle
Lift-and-shift reproduces data-centre waste at cloud rates — the discipline to come out ahead.
10 Jun 2026 · 18 min read
OpenTelemetry: instrument once, route anywhere
Separate instrumentation from destination and vendor lock-in stops setting the bill.
5 Jun 2026 · 17 min read
Microservices or a modular monolith? The question to ask first
The distributed-systems tax, itemised — and when microservices actually justify it.
28 May 2026 · 18 min read
Kubernetes cost optimisation: a utilisation problem, not a price problem
The average cluster uses about 10% of its CPU — fix sizing before touching pricing.
22 May 2026 · 21 min read
Cloud repatriation: when leaving the cloud actually pays
An engineering framework for whether leaving the cloud actually pays for your workload.
15 May 2026 · 15 min read
FinOps for DevOps teams: cost accountability without the blame
Make spend as observable as latency — visible, owned and continuously improved.
8 May 2026 · 16 min read
Curated collections
Scaling & reliability
Platforms that hold when demand doesn't arrive gradually.
- Autoscaling for traffic spikes: beyond a single HPALayer pod, node and event-driven scaling — a lone HPA won't survive launch day.21 min read
- Cell-based architecture: buying down blast radiusStop partial failures from spreading: blast-radius boundaries you can drain.20 min read
- SLOs and error budgets: turning reliability into a numberTurn “is it reliable enough?” from an argument into a number with a policy.18 min read
Cost engineering
Spend that falls without performance following it.
- Kubernetes cost optimisation: a utilisation problem, not a price problemThe average cluster uses about 10% of its CPU — fix sizing before touching pricing.21 min read
- Spot instances in production: 60% off without the 3 a.m. pages60–90% off on-demand without the 3 a.m. pages — engineering, not luck.21 min read
- Catching Cloud Cost Anomalies Before the InvoiceTreat spend as a real-time engineering signal, not a month-end finance surprise.19 min read
Delivery & security
Shipping faster while tightening the gate.
- GitOps in practice: your cluster should match your git historyGit as the source of truth, the cluster reconciling to it — and where that breaks.19 min read
- From scripts to pipelines: a CI/CD maturity modelYou don't need the most advanced pipeline — you need the next rung up.21 min read
- Securing the CI/CD supply chain: DevSecOps that doesn't slow you downYour pipeline is attack surface — controls that run inline, not gates teams skip.17 min read
Turn the reading into a diagnosis.
The Reality Check console applies these topics to your own platform — ten minutes, four instruments, a verdict.