UPS on bypass.

Trigger, action plan, and escalation criteria.

Related metric

ups_battery_readiness

Trigger

ups_mode == bypass

Value drivers

Operational continuity

Preconditions

The UPS operating-mode signal (online / battery / bypass) mapped over SNMP (UPS-MIB) or Modbus and verified per unit; which UPS protect critical load recorded so a bypass on a critical unit is prioritized.

Human role

AI detects the move to bypass, attributes it to the specific UPS, and routes the referral immediately; operators confirm the event is real, notify the customer, and coordinate the vendor to restore protection.

Action plan

(1) Confirm the bypass state is real and current, not a stale or mis-mapped mode point, before alerting — but treat a confirmed bypass as time-critical, not a cadence item. (2) Attribute it to the specific UPS and, where available, the reason (a maintenance bypass vs a fault forcing it), so the referral is precise. (3) The load is unprotected until the UPS is restored — notify the customer and the electrical / power vendor out of cycle; Keedian does not take the UPS off bypass. (4) Distinguish an authorized maintenance bypass (scoped, expected) from an unplanned bypass so a planned service window is not alarmed as an exposure.

Escalation

(1) A UPS on bypass protecting critical load → direct customer + vendor alert immediately; the bridge is gone until it is restored. (2) A UPS that stays on bypass past its expected maintenance window → escalate as an unplanned exposure. (3) Recurring unexplained bypass events on the same unit → flag a UPS fault for a deeper vendor engagement.

Prevention

(1) Verify the operating-mode signal maps correctly per UPS before arming so a mode-mapping fault is not read as bypass. (2) Record which UPS protect critical load so a bypass on a critical unit is surfaced first. (3) Log every bypass event with its reason so recurring unexplained bypasses surface a developing fault for the vendor.