Blog
Notes from building Stoicsoft. Tools, processes, and the occasional rant about software that wastes your time.
Showing 11–20 of 44 · Page 2 of 5
devops· Jun 3, 2026· 8 min readYour deploy says success, the old container is still serving traffic — close the trust gap
Operators lose trust when the deploy dashboard reports success but users still see the previous version. The fix is verify-after-deploy: confirm the live container is the new one before celebrating.
monitoring· Jun 3, 2026· 7 min readSimple cron heartbeats beat heavyweight monitoring for side projects
When developers ask 'how do I monitor my cron jobs?', the upvoted answers are ntfy, gotify, and healthcheck pings — not full monitoring stacks. Low-friction visibility wins for side projects.
devops· Jun 3, 2026· 8 min readTurn deployment runbooks into reusable skills — stop re-fighting the same eight gotchas
Every successful deploy hides a list of gotchas the operator figured out by hand. Capturing that flow as a reusable skill — not a doc — is the difference between tribal knowledge and a repeatable system.
engineering· May 25, 2026· 5 min readAI terminal sprawl: taming multi-repo coding CLI chaos
Running Claude Code, Codex, and shell sessions across three repos means twelve terminals, zero overview, and wrong-repo commands. Here's how to stop the sprawl before it stops you.
engineering· May 22, 2026· 4 min readThe CLI re-auth tax: scope your credentials per project, not per machine
Switching repos shouldn't mean re-authing five CLIs. The fix is scoping gh, gcloud, kubectl, aws, and ssh-agent state to each project — turning context-switch from an 8-minute parade into a cd.
engineering· May 22, 2026· 4 min readReusable project boot packs: the consulting onboarding pattern that beats rebuilding your shell
Every new client engagement starts with the same 40 minutes of CLI setup — wrong gh account, stale gcloud project, missing aliases. A project boot pack collapses that to one command and ships with the repo.
engineering· May 22, 2026· 4 min readPer-project handoff notes: the missing layer between your CLI history and your standup
Bouncing across three or four repos in a day, your terminal history isn't a handoff. The fix is per-project notes — auto-derived from the session, scoped to the repo, surfaced when the next engineer cd's in.
engineering· May 14, 2026· 7 min readTerminal session replay and blame trails: the layer pair-debugging tools keep skipping
Shared tmux and SSH sessions solve the live-collaboration half of pair-debugging. They don't solve the question that comes after the incident — who ran what, in what order, with what output. Replay and blame are the layer that closes the loop, and they're surprisingly cheap to add.
engineering· May 14, 2026· 7 min readTiered model routing: cheap-first, escalate on demand, and the AI IDE default nobody ships
AI IDE users burn premium credits on tasks a smaller model would have solved in two seconds. The fix isn't a smarter model — it's a router that runs the cheap tier by default and only escalates when the cheap tier flunks a verification step. Here's how that should be wired into the IDE, and what teams build today to approximate it.
engineering· May 14, 2026· 7 min readOne-click PDF prep packs: collapsing the freelance client deliverable workflow
Freelancers send PDFs to clients all day — invoices, contracts, scanned IDs, signed proofs — and the prep workflow looks the same every time: compress, redact, watermark, merge, sign. Today it's six tabs across three cloud tools. The win is a single local preset that runs the whole pack in one click, with the source files never leaving the desktop.