How to Escalate a Suspect PDU or Branch Circuit Without Guessing

"PDU is bad, please replace" is the escalation that bounces. Power faults love to blame the strip, and the strip is usually innocent — the C13 cord worked loose in a move, the branch circuit is overloaded, or a breaker tripped on load. Facilities knows this, so they move fast only when the ticket proves the fault follows the power path, not the server. Here is the isolation order and the note that gets it fixed.

Safety first. Everything below stops at the PDU outlets. Do not open electrical panels, do not reset a breaker more than once without facilities approval, and follow your site's SOP/MOP/EOP for any energized work. When in doubt, escalate earlier — a safe escalation beats a fast mistake.

Rule out everything else first (one variable at a time)

  1. Verify the ticket. Rack/U, which PDU (A-feed or B-feed), outlet numbers, and exactly what dropped vs what stayed up. If only A-side devices are down, you are probably looking at a feed or circuit problem, not a server problem.
  2. Read the PDU itself. Display or LEDs: amp draw per phase, outlet status, alarm indicators. Photograph or write down the readings before touching anything — post-touch readings prove nothing about the original state.
  3. Reseat the power cords. C13/C14 and C19/C20 connectors work loose after rack moves and maintenance. Push each cord firmly home at both the PDU and the device end. Record what changed after this alone — a surprising share of "dead PDUs" end here.
  4. Suspect load → known-good outlet, same PDU. Move the down device to a tested outlet on the same PDU. Powers up? The outlet (or the original cord) was the fault. Still dark? Keep going.
  5. Suspect load → opposite feed. Move it to the B-feed PDU (or A, if you started on B). Powers up now? The A-side path — PDU, branch circuit, or breaker — is carrying the fault, and the servers are cleared.
  6. Check the branch circuit. Breaker position (on vs tripped), panel schedule, and the circuit rating vs the PDU's amp readings. A 20 A circuit feeding a PDU drawing 19.5 A will trip on any spike — that is an overload escalation, not a hardware failure.

The numbers to capture before you escalate

An escalation note that gets answered

15:10 — Power down: srv-1042, srv-1043 rack C12 U18-U19, PDU-A outlets 7-8.
Both verified against ticket. PDU-A display: phase A 0.0 A, phase B 0.0 A,
outlet LEDs dark; PDU-B normal (phase A 6.1 A, phase B 5.8 A).
Reseat C13/C14 cords both ends: no change.
Test 1: srv-1042 to PDU-A outlet 12 (known-good) — still dark.
Test 2: srv-1042 to PDU-B outlet 4 — powers up, boots clean. Fault follows
the A-side power path; servers cleared.
Breaker for PDU-A branch circuit (panel E-4, 20 A): found TRIPPED. Reset once
per SOP — tripped again under load at 15:22.
Ask: facilities to inspect branch circuit / breaker for panel E-4, PDU-A
outlets 7-12; suspect overloaded circuit or failing breaker. Devices on
PDU-B for now. Next update 16:00 or on your go-ahead.

The mistakes that get escalations bounced

This plugs into the full floor workflow in RackReady Core ($19): the fixed-order no-power checklist, escalation and handover formats, and ticket templates. The free sample includes the universal break-fix checklist. No-power isolation in depth: server won't power on checklist; escalation format: how to write an escalation ticket.

See RackReady   Free sample   Ticket note examples

Support: support@hasherventures.com

Free tool: turn your checks into a structured ticket note with the RackReady Ticket Builder.