When a device fails, there are two ways to get it working again. An engineer attends site with a replacement part and repairs it in place, or the unit is shipped to a repair depot and sent back once fixed. The right choice depends on the device, the site and what the business loses while the device is away.
Start with downtime tolerance
The first question is how long the organisation can operate without the device. A core switch or a storage controller usually has no acceptable absence. A label printer in a back office, or a monitor with spares on the shelf, can often be gone for a week without anyone noticing.
Depot return adds transit time in both directions, plus queue time at the depot. If the device has no spare and its failure stops work, on-site repair is usually the only option.
Spares decide what an engineer can do
An engineer on site is only as useful as the parts they carry. On-site repair depends on the right field-replaceable unit being stocked near the site, at a compatible firmware revision. The common failure mode is the second visit: the engineer confirms the fault, orders the part and leaves, and the device stays down until the part arrives.
A depot usually holds deeper stock and bench equipment for component-level work. That matters for older equipment, where parts are scarce and out-of-warranty break-fix may only be possible at a bench. Ask any provider where the spares for each device model are held.
Data-bearing devices may not be allowed to leave
Many organisations do not allow storage media to leave the building while it holds data. That covers drives in servers and storage arrays, and less obvious items: multifunction printers with internal disks, and network appliances that keep logs or configuration.
Where that policy applies, a depot return needs the media removed and kept on site, or wiped first with a record to show it was done. Both steps need someone on site. Some hardware support contracts, including options sold by vendors such as HPE and Dell, allow a failed drive to be kept rather than returned. Check yours before the first failure.
Count the full cost, including logistics
Depot repair often looks cheaper because no engineer travels. The comparison changes once logistics are counted. Someone on site still has to power down, de-rack, pack the device and book the collection. A loan unit may need shipping out, installing and swapping back. Cross-border shipments add customs paperwork, duties and delay.
On-site repair carries travel time, out-of-hours rates and the cost of holding spares near each site. For a site near a service centre, that may be modest. Across many countries, it may not be.
Decide per device type, not per estate
Most organisations use both. Classify each device type against three questions:
- Does its failure stop work, with no spare to cover the gap?
- Does it hold data that policy says cannot leave the site?
- Are parts for it held near the site?
A yes to either of the first two, or a no to the third, points to on-site repair. Record the answer in the service schedule for each device type, so the decision is already made when the fault is logged at two in the morning.
