Every monitor reports a proxy, and the proxy is not the thing

What this site is

Readout covers tools for monitoring systems and tools for monitoring people. Those are two markets with different buyers and different vocabularies, and they are here together because they rest on one act: choosing something countable to stand in for something that matters.

A ping does not measure that the service works. Processor load does not measure that anyone was served in time. Hours logged and keys pressed do not measure that the work got done. In every case the number is real, the measurement is honest, and the thing anybody cared about was never observed at all.

The question this site asks of every tool is therefore how small that substitution is, and what decision the tool lets you make that you cannot make today. Where the answer is nothing, the tool is furniture that costs money and attention.

How the selections work

Fourteen of the forty-two pieces here are selections, and none of them is a ranking. Each begins with a situation, names the class of problem, sets out the criteria that decide the outcome, and only then mentions options. A product chosen before the class is a coin toss with a purchase order attached, because the classes solve unrelated problems and a tool bought across a class boundary works exactly as designed while failing the job it was bought for.

Every option carries a line describing what it costs you, including the ones we think are sensible. That line is the reason the rest is worth reading.

What is deliberately absent

No scores out of ten. A score implies a measurement, and we have not made one. No claim to have tested anything we have not tested. No verdict naming one best tool, because the class decides the outcome more than any feature list does. And no prices reproduced without the date they were read and the source they came from, in a market where published rates change quarterly.

In the workforce half, three further limits apply. Legal material names its jurisdiction, stays general, and is not advice. Nothing here describes how to monitor people without their knowledge, how to avoid a notification duty, or how to make capture difficult to detect. And every article ends with a section stating what it cannot verify, including where this site has an interest of its own.

Who it is for

An engineer deciding whether a cluster needs its own monitoring or whether the provider's console is enough. A manager handed a mandate from a board and working out how much of it can be scoped down. A finance lead reading an invoice that doubled while traffic did not. A team lead who has been asked to install something their team will hate, and would rather understand it before the conversation.

None of them is being taught their job. What is supplied is the part of the picture that sits in another department, in a form that survives being repeated to a colleague who was not in the room. That is why every piece ends with six numbered points: they are the version that gets carried into a meeting.

The thread that connects the two halves

Everything on this site is one mechanism appearing in different clothes. A threshold is a classifier that is wrong in two directions at once, and so is an idle-time setting. Alert fatigue is a response to a signal that is usually wrong, and so is the shrug a manager gives a productivity dashboard. A baseline taken during a bad fortnight makes overwork normal, and so does one taken during a crunch.

The sharpest version is the oldest observation in the subject: attach a consequence to a measure and it will be improved directly, by routes that are cheaper than doing the work. That is why a market exists in devices whose only function is to move a pointer, and why an availability figure used to compare teams produces narrower definitions of incidents rather than fewer of them.

Neither outcome requires anybody to behave badly. Both are what the arrangement selects for, which is the argument for designing the arrangement deliberately.

Where to start

The first section is the foundation and the other five depend on it: proxies, baselines, thresholds, retention, audiences, and what happens once a measure is used to judge. If you are fighting a bill, the infrastructure section has the arithmetic. If you have been told to install something, the workforce selections were written for exactly that position.