The Waffle House Index and Other Ways to Measure Civilisational stress

There are many ways to measure disaster. Meteorologists have wind speeds, seismologists have magnitudes, economists have charts that go down and to the right in an unsettling shade of red. None of these, however, answer the question most people actually ask during a crisis, which is: can I still function as a human being? The genius of the Waffle House Index is that it ignores the physics, skips the modelling, and goes straight to the point. If the Waffle House is open, civilisation remains on speaking terms with itself. If it is serving a limited menu, things are uncomfortable but survivable. If it is closed, then you should stop pretending this is “a bit of bad weather” and start acting accordingly.

This is not science in the academic sense. It is something better: operational truth. Waffle House does not close lightly. It has generators, supply chains, disaster playbooks and a corporate culture apparently fuelled by caffeine, spite and a refusal to accept the concept of “impossible”. When even they decide to shut the doors, the problem is no longer abstract. It is real, local, and has probably ruined your day.

The internet, for all its blinking lights and heroic self-regard, lacks such an index. We measure it with latency graphs, uptime percentages, packet loss, availability zones and other terms that sound reassuring right up until you realise none of them tell you whether you can pay for petrol, email your boss, or convince a confused relative that the world has not, in fact, ended. When the internet goes wrong, the official metrics often insist everything is fine. The public, meanwhile, is screaming into the void, assuming it can still resolve DNS.

What we need is an equivalent of the Waffle House Index for internet outages: a blunt, human-centred measure of whether connectivity still performs its basic civilisational duties. Not whether it is theoretically reachable, not whether a dashboard somewhere is green, but whether life can proceed without improvisation, profanity or a sudden interest in candles.

The Myth of “Uptime”

The modern internet is obsessed with uptime. Ninety-nine point nine per cent, four nines, five nines, enough nines to staff a small convent. These figures are presented as guarantees of reliability, but they are really accounting tricks. They assume that outages are neat, bounded events that politely announce themselves and affect everyone equally. In reality, most internet failures are partial, asymmetrical, and deeply unfair.

Your ISP may be “up” while DNS is having a nervous breakdown. Your cloud provider may be “operational” while authentication is unavailable, which is a bit like saying a bank is open but has lost the concept of identity. Payment networks may technically function while transactions fail just often enough to erode your will to live. From the perspective of a status page, this is fine. From the perspective of a person trying to exist in 2026, it is not.

The problem is that internet metrics tend to measure infrastructure health, not societal usability. They tell engineers whether packets are moving, not whether people can do anything meaningful with them. The Waffle House Index works because it inverts that logic. It asks not “is the system alive?” but “can it perform its purpose under stress?” The purpose of the internet, despite occasional claims to the contrary, is not to route packets efficiently. It is to allow modern society to function without reverting to carrier pigeons and shouting.

Choosing the Internet’s Waffle Houses

Every Waffle House Index needs a Waffle House, and the internet, regrettably, does not have a single diner chain staffed by indestructible short-order cooks. What it does have are anchor services: systems so deeply embedded in daily life that their failure is immediately felt, regardless of how healthy the underlying network claims to be.

These services share certain characteristics. They are normally reliable, boring in the way only essential infrastructure can be, and taken for granted right up until they stop working. Public DNS resolvers, for example, are rarely discussed at dinner parties, but when they fail, nothing else matters. Payment networks are similar. You can survive a surprising amount of technological inconvenience until you cannot pay for food or fuel, at which point philosophical questions about decentralisation lose their charm.

Cloud control planes occupy an especially bleak corner of this landscape. When they fail, the servers may still be running, the data may still exist, and the lights may still blink reassuringly, but no one can touch anything. This is the digital equivalent of being locked out of your own house while watching the television through the window.

Email, authentication systems, mobile data networks, and a handful of retail and logistics platforms round out this grim pantheon. None of these are glamorous. All of them are necessary. Together, they form a reasonable approximation of the internet’s Waffle Houses: not perfect, not universal, but close enough to reveal when things have gone properly wrong.

Green, Yellow, Red: The Comfort of Simplicity

The beauty of the original Waffle House Index lies in its simplicity. It does not offer nuance. It does not apologise. It gives you three colours and expects you to cope. Any internet equivalent should do the same, because complexity is the enemy of clarity, and clarity is the only thing anyone wants during an outage.

A green state is boring, which is precisely the point. Services work. Transactions go through. Authentication succeeds. The internet fades back into its natural role as an invisible utility, and everyone forgets to appreciate it. A yellow state is where the drama begins. Things technically function, but only just. Pages load eventually. Payments fail often enough to induce paranoia. Support desks begin using phrases like “intermittent issues”, which is corporate for “we have no idea”.

Red is mercifully unambiguous. Critical anchor services are unavailable. Not slow, not flaky, but gone. DNS does not resolve. Payments do not clear. Mobile data is absent or ornamental. At this point, arguments about whether the internet is “really down” become academic. From a functional perspective, it is closed, and someone should put the kettle on.

This colour system does not replace technical diagnostics. It ignores root cause, architecture, and blame. That is not a flaw. It is a feature. The Waffle House Index does not care why the storm happened; it cares whether you can get breakfast. An Internet Connectivity Functionality Index should show the same lack of curiosity.

Measuring Impact Without Pretending to Be Clever

One of the most dangerous instincts in technology is the urge to be clever when clarity will do. An internet Waffle House Index does not need machine learning, predictive analytics, or a blockchain, and the fact that someone will propose all three should be treated as a warning sign. What it needs is corroboration across independent signals.

Public status pages are a start, though they must be treated with scepticism, as they are written by people whose incentives do not always align with reality. User-reported outage aggregators add noise, but noise is useful when it comes from many mouths shouting the same thing. Active measurement platforms can confirm whether services are reachable from multiple vantage points. Payment failure rates, authentication error spikes, and mobile network availability reports complete the picture.

None of these sources is authoritative alone. Together, they provide something more valuable: consensus. When enough independent indicators agree that essential services are degraded or unavailable, the index changes colour. No debate. No press release. Just a quiet acknowledgement that things are not normal.

Why This Makes People Uncomfortable

There is a reason such an index does not already exist, and it is not technical difficulty. It is that impact-based measures are deeply uncomfortable for organisations that prefer to control the narrative. A system can be technically “up” while being functionally useless, and an index that says so removes the fig leaf.

Engineers dislike it because it abstracts away the elegance of their systems. Executives dislike it because it reframes outages in human terms. Regulators dislike it because it highlights dependencies they would rather not explain. Everyone, in short, dislikes being told that despite the dashboards, the diner is closed.

This is precisely why it would be useful. The Waffle House Index became famous not because it was official, but because it was honest. It spoke in a language that made sense to people who did not care about wind shear or pressure gradients. An internet equivalent would do the same, translating a labyrinth of protocols and providers into a single, blunt question: can society still operate online today?

Civilisation as a Service

At its core, the proposal for an Internet Waffle House Index is an admission of how dependent we have become. We no longer use the internet merely for entertainment or convenience. We use it to authenticate ourselves, to exchange value, to coordinate work, to access government, to reassure loved ones that we are still alive. When it fails, the effects cascade far beyond the screen.

Traditional network metrics struggle to capture this because they were not designed to. They emerged from a world where connectivity was optional and failure was acceptable. Today, connectivity is infrastructure, and infrastructure must be judged by whether it fulfils its social role under stress.

The Waffle House Index understands this intuitively. It does not care about optimal performance, only about continued operation. It assumes degradation will happen and asks whether the system adapts or collapses. Applying that mindset to the internet is less radical than it sounds. It is merely an acknowledgement that packets are not the point. People are.

When the Internet Is “Open, Limited Menu”

Perhaps the most interesting state is yellow, because it is where most real outages live. Total failure is rare and dramatic. Partial failure is constant and exhausting. In a yellow internet state, people can work, but slowly. They can pay, but nervously. They can communicate, but not reliably. Productivity drops, tempers shorten, and everyone suspects their own setup first.

This is the digital equivalent of a Waffle House serving toast and coffee because the grill is down. It is still open, but no one would pretend this is fine. Recognising this state formally matters because it validates experience. It tells users they are not imagining things, and it tells operators that degradation counts, even if the lights are still on.

Ignoring yellow states is how trust erodes. People remember being told everything was operational while nothing worked properly. An index that acknowledges degradation without melodrama offers a rare thing in outage communication: honesty.

A Modest Proposal, By Internet Standards

No index will fix the internet. This one will not prevent outages, eliminate single points of failure, or stop engineers deploying on Fridays. What it can do is provide a shared language for impact, one that cuts through the comforting lies of uptime and availability.

The Waffle House Index endures because it is humble. It does not claim to measure everything. It measures one thing well: whether a resilient, ordinary service can operate. An Internet Connectivity Functionality Index would do the same, replacing technical bravado with practical truth.

When the index is green, we can all relax and go back to ignoring the infrastructure that sustains us. When it turns yellow, we know to lower expectations and show patience. When it turns red, we stop pretending this is a minor inconvenience and start treating it like what it is: a disruption to the basic operating system of modern life.

And if nothing else, it would give us a simple sentence to say when the internet collapses in some novel and imaginative way. Not “the BGP routes flapped due to a misconfiguration”, not “we are experiencing elevated error rates”, but something far more honest.

The diner is closed.