When you’re the only dev and holidays depend on new hires
Description
Three-panel Pablo Escobar "waiting" meme. Panel 1 (top): the character sits alone on a white porch swing in an empty yard; large white caption reads "WAITING FOR YOUR COMPANY TO". Panel 2 (bottom-left): he stands motionless in a dim kitchen lit by a single bulb; caption over this panel says "HIRE A SECOND DEVELOPER". Panel 3 (bottom-right): he stands beside an empty swimming pool, staring into space; caption reads "SO YOU CAN GO ON HOLIDAY". The image humorously captures the bus-factor risk and work-life imbalance of single-developer teams who can’t take time off until management finally hires backup coverage
Comments
7Comment deleted
We’ve got 5 replicas per microservice for “high availability,” but the org is still running the entire engineering team as a single-replica deployment - waiting for the HPA to scale humans before I dare file PTO
The same company that won't pay for redundancy in infrastructure definitely won't pay for redundancy in humans
A bus factor of one means HR doesn't need a vacation policy - just a key-person insurance policy and your personal cell number
Ah yes, the classic single-threaded organization pattern where your vacation request enters an infinite blocking call waiting for management to approve hiring budget. Meanwhile, you're the only one holding the production pager, all the tribal knowledge, and the SSH keys to that one legacy server nobody else knows exists. The bus factor is 1, the deployment pipeline is 'you', and your PTO balance grows like technical debt - accumulating interest but never getting paid down. At least Pablo had a cartel to delegate to; you've got a backlog, three critical systems, and a Slack channel full of 'quick questions' that somehow always arrive at 4:45 PM on Friday
Infra: multi-region, self-healing, five-nines; org chart: bus factor 1 - my PTO fails Raft because there’s no quorum when I’m on holiday
Unlimited PTO, but we’re a single‑replica deployment: my vacation keeps failing admission because the PodDisruptionBudget has maxUnavailable=0 and the headcount HPA is set to 0
The singleton developer pattern: globally accessible, but hell to instantiate a second instance