"Nothing is wrong" and "nothing is running" keep coming out as the same value

I have been commenting on launches here most days since the end of July. One argument has come up in about a dozen of them without me planning it, and twice the maker went and found the same defect in their own code afterwards. That seems worth writing down properly rather than repeating in other people's threads.

The shape is always the same. A system has a value that means healthy. It usually does not have a separate value for "I could not tell". So the second collapses into the first, and it collapses in the direction that looks fine.

The clearest version came from mectrics, a Mac monitoring app. I asked Faruk what his CLI's check command exits when it cannot read the sensor at all. If that is 0, then "within limits" and "I could not look" are the same exit code, and a cron job sits green while the machine cooks. He had already handled that one. What came out of it afterwards was better: a watch could evaluate a cached reading after the provider underneath it had already failed, and report healthy from it. That is not silence. That is a green result computed from a reading that no longer existed.

Aadil found the same collapse in Pushary one layer up. A session that exited cleanly and a session swept for going stale both write offline, and the board files offline under done. His reframing is the sharpest thing anyone has said to me about this: it is an encoding problem, not a monitoring one. The distinction the emitter throws away cannot be recovered downstream, no matter what you build down there.

Which is why "be more careful" does not fix it. Writing offline is the locally correct choice every single time someone makes it. Nobody is being lazy. The only defence I know of is making the collapse inexpressible, so the value for nothing has to carry a reason with it.

Three cheap traps that follow from the same thing:

A health check tells you the process is alive, not that the work moved. Stuck on the same step for six hours is healthy. Waiting on a webhook that will never fire is healthy. Last progress is the number to put next to last heartbeat.

A daily digest of what the system did comes from the same system. If it dies at 3am the digest dies with it, and a missing digest reads as a quiet Tuesday. The count, including zero, has to be emitted by something else.

Greying out a stale number does not help either. Grey reads as quiet, and quiet is the exact meaning you were trying not to send.

So, answerable without going and looking at anything: which value in your system currently means both "fine" and "I did not check"? And if you have built the three state version, what did it actually cost you, because I suspect the honest answer is that it is boring and nobody thanks you for it.