REIN can be tricky to be find, and determine. There are all sorts of 'Interference' noises that can be heard on a radio/official 444B tester, but to be absolutely 100% certain, a spectrum analyser should be connected to the circuit at the far end (i.e: the actual premises) to see if the 'noise' is actually drowning out the signal ??
We get instances whereby there is extremely low-level 'noise', not just REIN but also crosstalk, that affects the circuit because the MPF's integrity ( the pair of wires from the Exchange to the premises) isn't as it should be.
Before a REIN case can be raised by a bog-standard engineer, the circuit must pass a PQT (Pair Quality Test) to a certain standard. As well as being free of any kind of 'Contact fault' (i.e.: low resistance between other working circuits, or actual earth), the individual leg-balance of each wire should be within 2-3 ohms of one another, and an AC Balance of -50dB or less should be obtained. Bear in mind the reading is in the 'negative', so ideally -60dB is better than -50dB, and -70dB is better than -60dB ..... etc etc.
As well as ensuring a good MPF, engineers have to install an RF3 Filter and an integrated filter also, to 'soak up' any minimal 'Noise' affecting issues.
The above is our (Openreach's) mandate, as once we have proven our MPF to be fault-free and that it meets SIN349 standards, then our remit is covered. However, being the caring company that we are, we do acknowledge that REIN is a problem and when workstacks are at a manageable level, we will generally assist the EU/ISP in locating the issue. Some are awful to locate, some are dead easy. Turning the electric off at the mains is acceptable if A) You feel competent, and B) It's non-problematic. By that, I mean that A) There are still some uninsulated live-busbar consumer units out there that could be lethal to the unassuming, and B) Turning consumer units off willy-nilly in a industrial environment can cause havoc.
Asbo has commented as to whether you can hear audible noise on the telephony side of the broadband circuit ?? (Assuming it's not a data-only circuit ??). This is a very simple step to take and if audible noise can be heard (crackling, hissing, humming) then I would suggest the MPF's integrity has collapsed in some way. Testing for that would be the best and easiest option in the first instance.
Regarding your query about the cable run around the warehouse ?? Yes, we are responsible for it with regard to 'Fair wear and tear' up to and including the NTE5 (socket). Again, by that I mean if you have decided to install an electric heater under the wire, or the company cat like to lag on the socket, then obviously charges (TRC) will apply. But, as I've said, we are not responsible for REIN faulting along any of the actual MPF. We do it as a gratis service.
I believe moves are afoot (and have been for a little while) for some kind of input financially from the ISP's when this kind of fault (REIN) arises. As the larger bandwiths are pushed down our cables, the more susceptible they become to REIN.
I'm not saying in any way, shape or form that your problem isn't REIN related, I just wanted to give a slight bit of guidance as to OR's responsibilities in regard to it, and that something like a very simple 'split-pair' on the MPF will give similar WHOOSH (DLM) readings, as REIN would (see DLM graph above posted up by Toblerone).
WHOOSH (Testing tool for DSL and other broad/wideband circuits) is a great tool for us when locating timings of REIN, especially the 'colour chart' in RRT (Retrospective Repair Tool). On Industrial Estates, on WHOOSH we can generally see a pattern with REIN from Mon-Fri (everything coloured in red), then weekends will see everything coloured in green. This is a massive indicator of REIN, coupled with good old knocking on other premises doors to ask if they are being affected, and driving around in the van with the radio set at 612Khz.
Sorry for prattling on, but that's my take on it.