Resources
The library
Everything is organised by topic. Entries marked in progress are queued and will fill in — nothing here is a dead end.
Cloud Infrastructure
3 entriesProviders, architecture, networking and what actually drives the bill.
Choosing between GCP, AWS and Azure for a small business
The provider differences that matter at 10 servers, and the ones that only matter at 10,000. A decision guide without the marketing.
What actually drives a cloud bill up
Egress, idle capacity, over-provisioned managed services and storage tiers. The line items worth attacking first.
Infrastructure as code when you are one person
You do not need a platform team to stop clicking in the console. A pragmatic starting point with Terraform.
Networking & DNS
3 entriesRecords, TLS, routing and how to debug something that will not resolve.
DNS records explained without the hand-waving
A, AAAA, CNAME, MX, TXT, and the ones you only meet when something is broken. What each record actually does and when to use it.
A network troubleshooting checklist that fits on one page
From the physical layer upward, in the order that finds faults fastest.
TLS certificates and why renewal is always the outage
Certificate expiry is the most predictable outage in existence. Automate it, and monitor the thing that does the automating.
IoT & Devices
3 entriesSensors, gateways, MQTT and getting telemetry somewhere useful.
Designing MQTT topic hierarchies that survive growth
Topic structure is the schema of an IoT deployment. Getting it wrong is expensive later, and most of the cost comes from wildcards.
Choosing an IoT gateway without overbuying
Protocol support, local processing, connectivity fallback and power. What the spec sheet does not tell you.
From a sensor on a desk to a dashboard you trust
The full path: device, gateway, transport, storage, visualisation, and where each one usually fails.
Automation & Scripting
2 entriesShell, Python, scheduled jobs and replacing manual work.
Writing scheduled jobs that survive contact with production
Locking, retries, idempotency and alerting on failure. The four things every cron job eventually needs.
Shell scripting habits worth having
Strict mode, quoting, traps and the handful of conventions that stop scripts from being dangerous.
Security & Access
3 entriesIdentity, least privilege, backups and what to do when it goes wrong.
Least privilege in practice, not in principle
Everyone agrees with least privilege and almost nobody implements it. A concrete way to get from admin-everywhere to scoped access without stopping work.
A backup you have not restored is not a backup
Retention, isolation, and scheduling restore tests so the first restore is not during an incident.
Incident response for a team of one to five
A deliberately short process: who decides, how you communicate, and what happens the day after.
Data & Reporting
2 entriesPipelines, modelling, storage and dashboards people trust.
Reporting numbers that are not quietly lying to you
Timezones, partial periods, double counting and the denominators nobody checks.
Do you need a data warehouse, or just a well-indexed Postgres?
The volume and query-pattern thresholds where the answer changes, and how to tell which side you are on.
Linux & Servers
2 entriesServices, permissions, logs and process management.
Linux permissions, ownership and the sticky bit
Unix permissions stop being confusing once you separate ownership from mode, and stop trusting octal memorisation.
Running a service properly with systemd
Unit files, restart policy, journald and resource limits — the parts that matter outside a container.
Web & Hosting
2 entriesDomains, deploys, CDNs and static hosting.
20 entries across 8 topics.