Manage system vital signs prompt

초코송이단 2025년 06월 07일

2,235

"The server is dead, why didn't anyone notice?" It's the most dreaded call you get in the middle of the night. Your users are already complaining because they can't access the service, and your monitoring system hasn't alerted you to anything. It's a developer's nightmare.

My team once had an incident where our payment server was down for two hours and no one knew about it, only to find out later from a customer complaint. It was a painful realization that "the server is alive" doesn't just mean the process is running.

Prompt.

복사

### System Health Check Architect Mode

[ Classify what to monitor ]

* Infrastructure Level

- CPU/Memory/Disk Utilization

- Network connectivity status

- Server response time

* Application level

- API endpoint responses

- Database connectivity

- External service dependencies

* Business logic level

- Core functionality working or not

- Data consistency validation

- User flow integrity

[ Design checks ]

>> Basic health check

GET /health → 200 OK

Response time: < 500ms

>> Detailed health check

GET /health/detailed

{ "status".

"status": "healthy",

"database": "connected",

"redis": "connected",

"external_api": "healthy"

}

>> Deep health check

Perform real business logic

End-to-end test scenarios.

[ Notification and response scheme ].

- Failure escalation

- Automatic recovery attempt logic

- Failure history management

Please design a perfect health check system for [service name].

Having such an organized health check system in place has given me a lot of peace of mind, as I'm able to catch signs of failure before they happen, and when they do, the average recovery time has gone from 20 minutes to 3 minutes.

Most importantly, my developers are much more relaxed and able to focus on their work because the system is telling them, "Wait, something's not right." Why not start checking your system's vital signs?

Previous Post List Next Post