The rack is leaking
A drip tray collects escaped coolant. It does not repair the leak or guarantee that every leak reaches its sensor.
Collection is only one layer
- Escaped liquid
Source still needs identification
- Tray and detector
Collection and sensing have coverage limits
- Identified alarm
Map sensor → rack → affected servers
- Qualified response
Protection, repair, and service validation
Know the parts
A cold plate transfers chip heat into coolant. A manifold distributes coolant. A compute tray holds computing hardware. A drip tray collects liquid where the design provides one.
A wet sensor identifies where liquid was detected, not necessarily where it escaped.
Connect the alarm to the work
Record the rack, sensor location, time, and related alarms. Facilities needs the physical scope; compute operations needs the affected servers and jobs.
Stopping a job does not make electrical equipment safe. Follow approved protection procedures and qualified response.
After repair, verify monitoring and service readiness. “Floor dry” is not the whole recovery test.
Try it yourself
A leak alarm names Rack C7, but inventory contains only server serial numbers. What is missing?
Check your answer
The mapping between sensor, cooling branch, servers, and workloads. Reconcile it rather than guessing. Use a paper coverage sketch; never pour liquid into equipment as a learning exercise.