What Really Happens When Your Fibre Line Breaks (and How to Get Back Online Faster)

At 13:01 on 24 July 2026, fibre network operator Octotel logged a fault affecting 47 streets in Strand, Western Cape - a residential outage notice later confirmed by reseller Cool Ideas listed the affected roads by name, from 1st Street through to Ash Close and dozens more. Octotel technicians were on-site assessing the damage by 13:42, began fibre resplicing at 13:51, and had the estimated time to resolution moving from an initial five hours down to three, then to 90 minutes, before confirming service fully restored at 17:18. Total outage window: a little over four hours, for a fault that had nothing to do with load shedding, and nothing wrong at any individual customer's house.

It's a useful, concrete example precisely because it's ordinary. Physical fibre breaks like this happen across every network in South Africa on a regular basis - from construction work clipping a duct, to cable theft, to ageing infrastructure failing outright - and they're handled in broadly the same way every time. Understanding that process is the difference between panicking, phoning support and waiting on hold, versus checking one page, confirming it's a known fault, and getting on with your day until it's fixed.

This piece walks through what actually happens between a cable being damaged and your connection coming back - who does the work, how long it realistically takes, how to tell a network-side fault from a problem in your own house, and what (if anything) you're entitled to if it happens to you.

It's worth being precise about the distinction, because the two failure modes get conflated constantly. Our piece on backup power for fibre covers what happens when the mains electricity feeding your ONT and router cuts out - load shedding, a municipal fault, or cable theft at an electrical substation. In that scenario the fibre line itself is completely intact; it's the equipment at your end that's lost power, and a modest UPS solves it.

A physical fibre break is the opposite problem. The glass itself has been severed or damaged - whether by a mechanical digger, a cable thief cutting into a duct expecting copper and finding fibre instead, ageing sheathing failing, or simple wear at a joint - and no amount of battery backup at your house changes that. The light signal simply has nowhere to go. The only fix is a technician physically locating the break and re-splicing the fibre strands back together, strand by strand, which is a skilled, on-site job that takes real time regardless of how good your home setup is.

Fibre splicing is precision work, not a quick patch. Once a fault is reported - usually via automated monitoring at the FNO's network operations centre, sometimes via a flood of customer complaints to resellers first - a technician has to physically locate the break along the duct or aerial run, which on its own can take longer than the repair. In the Strand incident, the roughly 40-minute gap between the fault being logged (13:01) and technicians beginning on-site assessment (13:42) reflects that travel-and-locate step.

Once found, the technician strips back the cable sheathing, cleaves the fragile glass fibre strands cleanly, and fuses them back together using a fusion splicer - a machine that aligns two hair-thin glass cores to within microns and welds them with a precisely controlled electric arc. Each splice is then tested for signal loss before the enclosure is resealed. A single cable can carry dozens of individual fibre strands, and depending on how many were damaged, a technician may need to repeat this process many times over before every affected customer is back online - which is part of why initial repair-time estimates (Octotel's moved from five hours, to three, to 90 minutes as the extent of the damage became clearer) tend to shorten as the work progresses rather than stay fixed from the start.

Fibre-optic cable is physically tougher than it sounds - the glass core is protected by layers of kevlar-like strengthening fibres, gel-filled buffer tubes and a tough outer sheath specifically so it can survive being pulled through underground ducts or strung between poles. But none of that makes it indestructible, and a handful of causes account for the overwhelming majority of physical breaks reported across South African networks:

• Third-party construction damage. A contractor digging a trench for water, sewer or electrical work without first requesting an as-built map of what's already buried is one of the most common causes of an accidental fibre cut - the duct simply isn't where the digger expected it, or wasn't marked at all.
• Cable theft and vandalism. Thieves targeting what they believe is copper sometimes cut fibre cables by mistake, since the two can look similar from the outside before the sheath is opened. Fibre itself has essentially no scrap-metal resale value, but the damage to the line is identical either way.
• Weather and environmental wear. Storm damage to aerial fibre runs, flooding that erodes shallow duct cover, and simple UV/heat degradation of ageing sheathing over many years all contribute to a background rate of faults independent of any single event.
• Rodent damage. A less glamorous but genuinely common cause at the "last mile" - rodents gnawing through cable sheathing, particularly on above-ground drop cables closer to individual homes.

None of these causes are something an individual customer can prevent, which is exactly why the fix sits entirely on the network operator's side, and why the diagnostic steps below matter more than any troubleshooting you could do at your own router.

Before phoning anyone, the fastest diagnostic is usually built into the ONT sitting on your wall. Every fibre network operator's terminal has a specific light that indicates a fibre-side signal fault, as opposed to a power, router or in-house wiring problem:

If that light shows a fault, the problem is upstream of your house - on the network, not your router, your WiFi, or your ISP account - and no amount of restarting your own equipment will fix it. If the fault light is fine and only your WiFi or a specific device is misbehaving, the fault is local, and troubleshooting your own router and devices is the right next step (see our guide to diagnosing a slow fibre line for that side of things).

South African fibre works on a wholesale/retail split: a handful of fibre network operators (FNOs) - Openserve, Vumatel, Frogfoot, MetroFibre, Octotel and others - own and maintain the physical infrastructure, while dozens of internet service providers (Afrihost, Webafrica, Cool Ideas, RSAWEB, Vox, Mweb, Supersonic and more) resell access to that same infrastructure to consumers. When a cable breaks, it's almost always the FNO's technicians doing the physical repair, regardless of which ISP's logo is on your invoice - your ISP is typically relaying status updates it received from the network operator, not fixing the fault itself.

That matters for where you look first. If your specific ISP's own systems (billing, email, a misconfigured account) are at fault, only that ISP can fix it. But if the FNO serving your address has a physical break, every ISP reselling that same network in your street is affected identically - which is exactly why checking a network-level status page, rather than only your ISP's Twitter/X account or support line, is often the fastest way to confirm what's actually going on.

Most FNOs and larger ISPs run some form of public status page, and several independent, multi-network trackers exist precisely because so many ISPs sit on the same handful of underlying networks. Before queuing on a support line, it's worth checking:

• Your own ISP's status or network page - most major providers (Afrihost, Vox and others) publish one that flags known Openserve, Vumatel, Frogfoot and MetroFibre incidents in real time.
• The FNO's own status page, if your ISP's site doesn't specify which network you're on - Octotel, for example, publishes a public notice history covering exactly the kind of incident described above.
• Independent, cross-network outage trackers that aggregate incidents across several FNOs at once, useful when you're not certain which underlying network serves your address.

If a status page already shows an active, being-worked incident for your area, there's usually nothing more productive to do than wait - phoning support at that point mostly adds you to a queue for information the status page already gave you for free. If nothing is listed there but your fault light shows red, that's the moment to actually log a fault with your ISP, since it may not be a known, already-flagged incident yet.

It's also worth noting how much these status pages tend to evolve during a live incident, which the Strand fault illustrates well: the estimated time to resolution shifted three times in under two hours (five hours, then three, then 90 minutes) as technicians got a clearer picture of the damage on site. Treat an ETR on any status page as a moving estimate that improves in accuracy over time, not a fixed promise - checking back periodically rather than fixating on the first number you see will generally give you a more realistic picture of when you're actually likely to be back online.

South Africa doesn't have a blanket, industry-wide rule guaranteeing automatic rand-for-rand credit for every hour of downtime - what you're entitled to depends on your specific ISP's terms and how the outage is classified. Some providers offer pro-rata account credit on request for extended, verifiable outages; others treat brief network-level incidents (a few hours, resolved same day, affecting a whole suburb rather than one customer) as normal service variability rather than a billable failure. The practical approach: keep a record of when your connection dropped and was restored, and ask your ISP directly and in writing whether a credit applies - don't assume either way.

If your ISP doesn't resolve a complaint to your satisfaction, South African consumers have a formal escalation path through the Independent Communications Authority of South Africa (Icasa). The process requires lodging a complaint with your service provider first and obtaining a reference number, then allowing the provider 14 working days to resolve it - only after that window, or if the response is unsatisfactory, can the complaint be escalated directly to Icasa's Consumer Affairs unit. It's a slower route than a same-day network fault, but it exists specifically for cases where a provider isn't engaging on a legitimate, unresolved service complaint.

A handful of practical steps make a real difference to how disruptive a physical fibre fault is when it eventually happens on your line, since - as the Strand incident shows - even a well-run network with fast-responding technicians can still mean several hours offline:

• Keep a mobile data failover in reach. A phone hotspot or a router with a SIM-based 4G/5G failover mode won't match fibre speeds, but it keeps email, calls and basic work moving during a multi-hour outage - a meaningfully cheaper insurance policy than most people assume.
• Know your ONT's fault light before you need it. Looking it up mid-outage wastes time; a 30-second check today (using the table above, or your specific ISP's support page) means you can self-diagnose the moment something goes wrong.
• Bookmark your network's status page, not just your ISP's. If you don't know which FNO serves your address, your ISP's coverage-check tool or a recent invoice usually says.
• If outage frequency genuinely affects you (working from home, running a business, VoIP or security systems), factor an FNO's fault-resolution track record into your next provider choice - see our comparison of fibre ISPs in South Africa for how different networks and resellers stack up on support and reliability, not just headline price.
• Consider a dual-WAN router if continuity is genuinely business-critical. Some higher-end routers can hold two active connections - a fixed fibre line and a cellular backup - and fail over automatically within seconds of detecting a drop, rather than requiring you to manually switch to a phone hotspot. It's a meaningfully bigger outlay than a basic UPS, and only worth it if an hour of downtime has a real cost attached, but it addresses exactly the scenario a UPS can't: a fully powered router with genuinely no fibre signal to carry.
• Ask a prospective ISP about outage history before you sign up, not just price and speed. A provider or reseller that can tell you plainly which FNO serves a given address, how that network reports faults, and what your options are during an extended outage is generally a better sign than one that can only quote a speed and a rand figure.

The Strand fibre break resolved in a little over four hours - fast by the standards of a physical cable repair, and a reasonable benchmark for what "the system working as intended" actually looks like when a duct gets damaged. It's a different failure to the load-shedding and cable-theft power outages that a UPS solves, and it's fixed by a technician with a fusion splicer, not a battery. Knowing your ONT's fault light, checking a status page before you phone support, and understanding that your ISP is usually relaying someone else's repair timeline rather than causing the fault yourself won't make a cable break happen less often - but it will make the next one far less stressful to sit through.