Most organizations don’t measure reality. They measure performance against numbers they invented. That difference matters.

A metric is supposed to reveal truth. The moment the number becomes the goal, the truth is gone.

From that point on, you’re not improving systems. You’re managing appearances.

BoringOps exists because this failure mode is everywhere.

Metrics Are Not the Work

When teams say they are “data-driven”, what they usually mean is metric-driven.

Dashboards fill up.
Targets get hit.
The system gets louder.
The humans get tired.

Nothing important improves.

Metrics were never meant to be accomplishments. They were meant to be signals. Once the signal becomes the score, the system starts lying.

The Rule You Cannot Escape

If a metric can be gamed, it will be. If a metric can be misunderstood, it will be.

People optimize for what you measure because you told them to.

A Real Example of How Metrics Get Gamed

Early in my career, our team handled a steady stream of trivial requests.

Permission tweaks.
Config nudges.
Service restarts.

They took two or three minutes. Fast. Human. Efficient.

Then leadership decided everything needed a ticket.

Once a ticket exists, someone will measure it. That part is inevitable.

So I adapted.

I asked every requestor to open tickets for even the smallest task. My teammates didn’t bother.

The result?

Every quarter, I closed ten times more tickets than anyone else on the team.

Did I do more valuable work?
No.

Did the system get slower?
Yes. We added friction purely to satisfy a number.

Did I look indispensable?
Absolutely.

Did I game the system?
Of course I did.

That’s the trap.

The moment metrics detach from meaning, they stop describing reality and start rewarding behavior that looks good on paper.

The Three Ways Metrics Break Systems

Numbers Replace Outcomes

Reward throughput and work fragments.
Reward deployment count and quality leaks.
Reward cost control and technical debt compounds.

You didn’t get what you wanted.
You got exactly what you measured.

Symptoms Get Optimized

MTTR is not a fix.
Error budgets are not safety.
Cost curves are not architecture.

These numbers describe consequences, not causes.

Optimizing symptoms makes the system harder to understand and harder to fix.

Activity Gets Confused for Signal

More dashboards do not mean more clarity.
More alerts do not mean more control.

If your team is drowning in metrics, you are already blind.

What BoringOps Measures For

Boring infrastructure has a shape.

It is predictable, repeatable, quiet, and boring to operate.

The Nine Pillars define that shape. Metrics only exist to show whether you are moving toward it or drifting away.

If a metric does not reduce chaos, it adds it.

Consistency

If the system behaves differently week to week, nothing else matters.

Deployment frequency is noise.
Deployment stability is signal.

Uptime percentages are noise.
Operational silence is signal.

Consistency is the prerequisite for truth. When behavior drifts, every other measurement becomes unreliable.

Simplicity

If a metric needs explanation, it has already failed.

Measurement should collapse complexity, not narrate it.

BoringOps keeps the list short:

  • stability
  • predictability
  • recovery clarity
  • operator friction
  • architectural churn

Everything else is decoration.

Transparency

Green dashboards mean nothing if the humans are exhausted.

Truth shows up in alert volume, incident recurrence, escalation frequency, and on-call fatigue. But the goal is not to push those numbers down through sheer effort. The goal is to remove the causes of the noise.

A metric is only useful if it reflects the team’s lived experience.

Stewardship

If you celebrate heroics, emergency fixes, weekend saves, and “great recoveries,” you are rewarding broken design.

Heroics are not excellence. They are evidence of neglect.

Good metrics make heroics uncomfortable to defend. They expose the design flaws that forced someone into a cape. Stewardship means protecting the system from the pressure to optimize for the wrong numbers, including the pressure that comes from dashboards.

The Only Acceptable Purpose of Measurement

Every metric must justify itself by doing at least one of these:

  • reduce chaos
  • increase predictability
  • free human capacity
  • expose architectural truth
  • move the system toward boring

If it does none of these, delete it.

The End State

When metrics regain meaning, something predictable happens.

Dashboards shrink.
Incidents fade.
Teams stop gaming numbers.
Leadership starts asking better questions.

The system gets quieter.

If your dashboards look healthy and your team feels broken, your metrics are lying.