"The application is on a server"
- ✗
One server runs everything
Web server, app, and database on a single EC2 - one OOM event takes it all down
- ✗
Manual SSH deployments
Someone types git pull in production - and one typo is a production incident
- ✗
No monitoring
The team finds out about downtime when a user reports it, not before
- ✗
No staging environment
Changes are tested in production - every deploy is a calculated risk
- ✗
Infrastructure undocumented
Only one person knows how the server is configured - and they're on holiday