Blog
Practical reliability thinking for small software teams.
Clear guides, operating frameworks, technical explanations, and original research for teams that need to know when software fails and communicate what happens next.
Featured
Fajita is now open source
The Fajita uptime monitoring core is now open source under AGPL-3.0. Self-host the verification engine or use Fajita Cloud for managed operation.
Core topic areas
- Monitoring
How to check websites and APIs on a schedule, confirm failures before paging anyone, and keep monitoring signals honest.
- Incident Response
Writing public updates, separating internal notes from customer language, and closing incidents without improvising under pressure.
- Status Pages
What belongs on a customer-facing status page, how to name components, and how to keep the page current during maintenance.
- APIs and Webhooks
Health endpoints, status codes, JSON assertions, authenticated checks, and signed webhook delivery.
- Cron and Scheduled Jobs
Heartbeat monitoring for cron, backups, and scheduled tasks that fail silently when nobody is watching.
- Reliability Metrics
Uptime math, maintenance treatment, and the difference between a percentage and a promise.
- Founder Operations
The smallest reliability stack that still protects a solo founder or tiny team before and after launch.
Latest articles
Fajita is now open source
The Fajita uptime monitoring core is now open source under AGPL-3.0. Self-host the verification engine or use Fajita Cloud for managed operation.
The Minimum Reliability Stack for a Solo SaaS Founder
Four monitors, one alert channel, and one status page. The smallest reliability stack that still protects a solo SaaS before customers notice first.
How to Monitor an API Without Creating Alert Noise
Use the Confirm Ladder: probe, retry, verify, then open an incident. Keep API monitors useful without paging on every transient blip.
How to Write a Useful Incident Update
Use the FACT template: facts customers can verify, action you are taking, cadence for the next update, and a calm tone. Templates you can paste under pressure.
What Should Go on a Public Status Page?
Use the Customer-Visible Component Rule. Show what customers feel, hide internal plumbing, and keep maintenance honest.
Heartbeat Monitoring for Cron Jobs and Scheduled Tasks
Heartbeats catch jobs that never run. Learn grace periods, late versus missed pings, and how to secure heartbeat URLs without agents.
How to Calculate Uptime Correctly
Uptime percent is allowed downtime divided by the measurement window. Define maintenance, check gaps, and the Measurement Contract before you publish a number.
Why One Failed Check Should Not Mean Downtime
One failed probe is a blip until confirmed. Learn Blip Budget thinking, verification, and when a single failure should still page you.