Plurality is not a promise that one will answer

The inference people make

Three addresses, so surely at least one is up. It is a comfortable thought and it does not follow from anything. Publishing is an act that happened once. The state of a service is a thing that changes continuously and independently.

The three strings on this page are text. They will still be text if every service behind them stops answering. Nothing about them updates itself, and nothing about them is a commitment.

Why the three can fail together

The routes are separate but they are not independent in the way the word redundancy suggests. Three routes to one service depend on that service. They may also share infrastructure, an operator, a hosting arrangement and a set of decisions, none of which this site can see.

So the failures a set of addresses protects against are the narrow ones described in what redundancy buys. Anything that reaches the thing behind all three reaches all three at once, and the section on all three quiet is about exactly that case.

What the reader is entitled to conclude

From the existence of a set you may conclude that at some point somebody arranged more than one route and published them. That is the whole of the entitlement.

You may not conclude that any of them answers now, that they are maintained, that they will be updated, or that anybody will tell you when one is retired. Each of those would need evidence of a kind that a published list cannot carry, which is the argument of what a list proves.

The cost of reading it as an assurance

A reader who treats the set as a guarantee draws a strong conclusion when the guarantee appears to fail. If all three are supposed to work and none do, then the market must be gone, and the search for a replacement begins.

A reader who never took it as a guarantee reaches a quieter conclusion: something is wrong somewhere, possibly at their end, and none of it is settled yet. The second reader is right more often and is in a far better state to make decisions.

Why this site publishes no figure of any kind

It would be easy to put a number beside each string, and every number would be an invention, because this reference does not open the addresses it prints. There is no uptime percentage here, no coloured light and no time of checking, on any page.

The full argument for that is in why there is no uptime number. The short version is that a figure would answer a question nobody actually has, while implying an answer to the question they do have.

Holding the set without over reading it

The right posture is unexciting. The set is a convenience worth keeping and not a claim worth relying on. Hold all three, expect no announcements, and treat any change in the set as something you will find out about late and by accident.

That posture costs nothing when everything works and saves the reader from a bad conclusion on the day something does not.

What a promise would look like if one existed

It would have to come from the market, reach you through a channel you already trust, and say something about a period rather than a moment. None of those three conditions can be met, which is why no such promise exists anywhere in this subject.

What does exist is a published set, which is a fact about an arrangement rather than an assurance about a state. Readers who keep those two apart are not being pessimistic. They are describing the situation accurately, which is what makes their reactions to it proportionate.

What this study is not. It is not a suggestion that anything is wrong with any address, and this site has no information of that kind about any of them.