Every codebase has a cron job that one engineer quietly owns. The deploy script only they can run. The backup job that emails "you know who to ask" when it fails.
That's bus factor 1, and it usually surfaces at 3am on a weekend.
Here's the 10-minute audit we run on Fridays:
1. Find the load-bearing humans (2 min)
crontab -l | grep -iE "alex|sam|priya" # every hit is bus factor 1
grep -rn "TODO\|FIXME\|ask:" deploy/ cron/ | grep -v "team"
Any hit where a human name is the answer to "how does this work?" is a single point of failure.
2. Write the one-pagers (5 min)
For each hit, one paragraph:
- what the job does
- how to run it manually
- how to tell it worked
If a paragraph doesn't fit, the job is too complex — that's your refactor list.
3. Automate the two scariest (3 min today, ship this week)
Runbooks rot; scripts do not. Cron + idempotent bash + a heartbeat check. The heartbeat matters: a job that "runs" but produces garbage is worse than one that fails loudly.
We packaged the full checklist — scripts, one-pager templates, heartbeat checks — into two small kits:
- Ops Starter Kit — checklists + automation scripts
- Automation Starter Pack — runnable bash/cron templates
Run the audit Friday afternoon. Fix the top two before Monday. Your successor inherits a system, not a mystery.
What's the oldest bus-factor-1 job still running in production where you work?
Top comments (0)