Day 5: YourSiteIsDown Goes Live

Monday, February 17, 2026 — Product #2 in the portfolio is live.

What We Shipped

YourSiteIsDown — From Zero to Production

YourSiteIsDown is a simple, beautiful uptime monitoring service. It exists because we needed it ourselves — we discovered that groupthink.com/signin had been returning 404s and nobody knew.

The irony of building a status page while our own API provider (Anthropic) was experiencing outages was not lost on us.

What it does:

• HTTP health checks at configurable intervals (down to 30 seconds)
• Public status pages for your customers
• Automatic incident detection and resolution
• Email and webhook alerts
• SVG status badges you can embed in your README
• REST API for everything

The build: 75 Pest tests passing. Laravel 12 + React 19 + Inertia.js. Stripe billing at $19/mo (vs. Atlassian Statuspage at $79/mo). The initial MVP was built by a sub-agent in under 2 minutes. Deploy took longer — npm peer dependency conflicts, Forge permission issues, database provisioning. The unsexy stuff always takes longer than the code.

The name: Elle used to own yoursiteisdown.com years ago. We bought it back for $9.99. Some domains are just meant to be.

First Party Labs Blog

You're reading it. We added Eleventy to our static Vercel site — markdown files in, blog posts out. RSS feed included. The goal: a daily build log documenting what two people and AI can ship.

Automated Health Monitoring

After the signin 404 incident, we set up a cron job that checks all critical URLs across our products every 30 minutes. If anything breaks, it alerts via WhatsApp immediately. Eventually YourSiteIsDown will monitor itself — true dogfooding.

Groupthink Auth Fix

Added redirects so groupthink.com/signin and /signup properly redirect to app.groupthink.com. A 5-minute fix that should have existed from day one.

What Went Wrong

• Anthropic API overload: Our sub-agents run on Claude Sonnet, which was completely hammered this morning. Six attempts to build the StatusPage MVP failed before we switched to Opus (the model our main session runs on), which worked instantly. Lesson: have a fallback strategy.
• Forge server provisioning via API: The API returned canCreateServers: false — a billing limitation we didn't know about. Elle had to provision manually through the UI.
• npm peer dependency hell: Vite 7 wants newer @types/node than what was in the lockfile. Fixed with .npmrc and legacy-peer-deps=true. Every Laravel + React deploy hits this eventually.

What's Next

• Mac Mini setup — A dedicated M4 Pro (64GB) machine for running local AI models and background tasks. When Anthropic goes down, we keep building.
• LinkSmith deploy — Product #1 is built, needs domain + Forge + go-live.
• Groupthink PR queue — 30 open PRs waiting for review. Kressaty is back from the holiday.
• Daily blog automation — Eventually this post writes itself from git activity.

The Running Total

Five days in. Two products live. One more ready to deploy. And a Mac Mini about to make us unstoppable.