Internet Down: What to Do in the First 30 Minutes
Florian Hödl
Network Monitoring and Automation
In short: First check with a mobile phone, WiFi switched off, whether the problem is confined to your building or affects the whole region. Then report the fault with the site's line identifier, the time it started, and a callback number that does not run over the failed connection. Keep the business running over a second path while repair is under way. Afterwards, write down what happened, because the second outage on the same line is no longer a coincidence.
The first question decides everything after it
When nothing works, the first question is not "whose fault is this" but: is it us or the provider? While that stays open, one person calls the provider while another reboots the router, and neither of them knows whether they are doing the right thing.
You get the answer in two minutes, without tools:
- Mobile phone, WiFi off. If pages load over mobile data, it is your connection that is down, not the internet. It sounds trivial and it removes half the diagnostic path.
- Cable instead of WiFi. A laptop plugged straight into the router. If that works, WiFi is the problem and the line is fine. By some distance the most common case.
- Read the lights on the router. Almost every device distinguishes between "no link to the line" and "link up, but no data". The first is a case for the provider, the second is often a case for your own configuration.
- Look outside your building. Outage reporting sites and your provider's status page usually surface large-scale failures within minutes. A call to the business next door on the same provider is just as conclusive and often faster.
If points 1 and 4 coincide, mobile data works and the outage pages are filling up, stop hunting for faults inside the building. It is regional, and your job shifts from "repair" to "keep going and document".
Not every outage fails completely
One thing 11 August showed many businesses: an outage is rarely all or nothing. That afternoon, data traffic kept running on many connections while voice, TV and the operator's central services were gone. Anyone reporting "the internet is broken" described something other than what had actually failed, and landed in the wrong queue because of it.
So check these separately before you report:
- Data: does any website load that does not belong to your provider?
- Voice: does a call reach the landline number? Does an outbound call connect?
- Internal services: can you reach your own systems, point of sale, time tracking, inventory?
Three answers instead of one guess. That is the difference between a fault report that goes straight to the right team and one that travels across the whole hotline. Why this split happens technically is covered in the review of the A1 outage of 11 August.
The report that does not die on hold
Fault desks work tickets by completeness, not by insistence. Have these five details ready before you dial:
| Detail | Why it makes the difference | |
|---|---|---|
| Line identifier or customer number of the site | Without it every call starts with a search. With multiple sites, the company number alone is worthless. | |
| Time it started | Determines whether your report gets merged with a known major incident. | |
| Which services are affected | See above. Decides which team owns it. | |
| What you have already checked | Skips the standard questions and raises the ticket's priority. | |
| A callback number that does not run over the failed connection | The most common avoidable mistake. If the callback goes to the dead number, you lose an hour. |
Ask for the ticket number and note the time you reported it. Without those two, nobody can reconstruct later how long the repair actually took, and that is exactly what you need when it happens a second time.
Keeping the business running during the outage
What is possible here is decided before the outage, not during it. Even so, more usually works than the first shock suggests:
- Mobile hotspot carries email, messaging and cloud applications for a handful of desks. It does not carry a building with thirty people, and it does not carry card payments if the terminal is wired. Anyone who has tried it once knows on the day how far it reaches.
- Call forwarding to a mobile is the single most effective measure for how you look from outside, because customers are not left ringing into nothing. It does have to be configurable without reaching the customer portal over the failed connection. That is where it usually breaks down on the day.
- Card payment on many terminals has a mobile fallback path that must be enabled before you need it.
- A clear notice outward. One line on the website or the company profile ("we cannot be reached by phone at the moment, please write to ...") costs two minutes and prevents the impression that you are closed.
If a site is business critical, none of these replaces a second, physically separate path. What redundancy can and cannot do is covered under fixed line and 5G as site redundancy. The key point: a second contract with the same provider over the same line is not redundancy, it is a second invoice.
After the outage: the ten minutes that pay off
Once everything runs again, the temptation is to get straight back to work. Note this down instead:
- Date, start time, end time, ticket number
- Which services were affected
- What the provider named as the cause, and whether they named one at all
This is not bureaucracy, it is the only basis on which you can later do two things. First, recognise whether a pattern is forming. Three faults on the same line in one quarter is not bad luck, it is a procurement matter. Second, talk about compensation or contract changes, because without documented downtime any such conversation is one assertion against another. If you get nowhere with your provider, the Austrian regulator RTR runs a dispute resolution body for disagreements between customers and telecom providers.
When it stops being an isolated case
A connection that fails several times a quarter, or constantly flips between reachable and unreachable, is not a monitoring problem. More monitoring does not make it better. It needs repair or replacement.
The distinction matters because both cases feel identical on the day and need completely different answers. Some faults announce themselves measurably over days, others arrive with no warning at all. Which group is which, and what follows from that for redundancy and alerting, is the subject of the article on faults with lead time.
What we take on
For the businesses we look after, the steps above do not run at their end but at ours: the fault shows up in internet monitoring before the first call comes in, the report goes to the right carrier with complete details, and the repair is chased. What that means day to day is described under managed connectivity.
If you want to know how your sites actually stand after the last outage: let us talk.
Frequently asked questions
How do I tell whether the fault is mine or the provider's?
Test with a mobile phone with WiFi switched off, and connect a laptop by cable directly to the router. If it works over mobile data but not over cable, it is your connection or your equipment. If the outage reporting sites fill up at the same time, it is regional and you can stop hunting inside the building.
What do I need to give when reporting a fault?
The line identifier or customer number of the affected site, when the outage started, which services are affected, what you have already checked, and a callback number that does not run over the failed connection. Ask for the ticket number and note the time.
The phone is dead but the internet works. How can that be?
Voice and data transport run over separate systems at the provider, even when both arrive over the same line. If the central voice platform fails, data traffic is unaffected. This is exactly the pattern seen in the Austria-wide outage on 11 August 2026.
What can I bridge with until the line is back?
A mobile hotspot for individual desks, call forwarding to a mobile, a mobile fallback path on the card terminal, and a short notice outward. All four only work if they were set up and tested before the outage.
When should I start thinking about a second line?
As soon as one outage day costs you more than the redundancy costs per year, or as soon as faults repeat at the same site. What matters is that the second path is physically separate. A second contract over the same line does not help when it counts.