I Run a Business Alone. My Monitoring Stack Is One Folder.
No Datadog. No $29/month SaaS. No Slack alerts routed through three vendors. When you are the entire ops team, every layer you add is another thing that can break while you sleep.
Here is the stack that actually survived my first year solo — all of it fits in one folder on a $5 VPS or a Mac mini.
The Three Files That Matter
1. sensor.sh — the loop that cannot lie
A cron'd bash script that curls every money URL and validates the content, not the status code. A 200 with an empty body is the most expensive lie in solo ops.
2. ledger.txt — the append-only trail
One line per check: timestamp, what was checked, what was found. When something breaks at 3 AM (and it will), you reconstruct the timeline in 60 seconds instead of guessing. date -u >> ledger.txt is 90% of incident response when you have no team.
3. healer.sh — restart, verify, log
Dead process? Restart it. Then prove the restart worked (check the port, check the endpoint). Log the verdict. A healer that doesn't verify is a prayer, not a healer.
The Rule That Ties It Together
Every check writes a line. Every healing gets a verdict. Every page to yourself includes the last 20 ledger lines. That's it. That's the whole doctrine.
I ran this stack through: a dead payment webhook (caught in 5 minutes), a DNS expiry (caught the day before, not the day after), and a scraper that silently returned empty JSON for two weeks (caught because the sensor validated content).
The Full Kit
I packaged the exact scripts — sensors, ledger, healer, pager test, escalation ladder — as Ops Starter Kit, plus Automation Starter Pack (monitoring/restart patterns) and Agent Ops 24/7 (runbook for AI agents that run your business unattended).
Everything lives here: HIVE80LAB — https://hive80lab.gumroad.com
Steal the architecture even if you never buy a thing: one folder, three files, zero dependencies. It beats a dashboard you forgot to check.
Top comments (0)