Software runs the world. It’s all around us, and it’s inescapable.

Bad software can leave us stranded in an airport. Evil software can topple governments. Good software can help save lives. Whatever you do, and wherever you are, software is more essential than we can fathom.

As software became ubiquitous, it’s practitioners grew manifold. Talented engineers worked really hard to abstract complex code into simple clickable buttons for users. With a simple click of a button, you could get food delivered, and cars navigate themselves.

Software’s outward simplicity and inward complexity attracted hundreds of people who wanted to build products that helped advance our lives.

This prolific growth meant lots of servers, systems, and complex backend machines that powered millions of applications that made our lives easier.

As more practitioners entered the fray, newer kinds of software emerged. We could do more with less. We could accommodate more engineers and build things faster, better. In the process, the underlying infrastructure got so massive, it was getting harder to identify where software was not performing as intended.

This sprawling infrastructure of machines were not resilient to failures. After all, human cognition and interpretation is varied and subjective. The language and emotion I have writing this might not be interpreted by you, the reader, in the exact 100% emotion I want to convey. We’re whimsical, subjective beings, and it’s what makes us unique.

But for the world of software, this meant things could fail. And when it did, it left a stinky mess. There was only one solution. The performance of software had to be measured, and acted on. Without the measurement of software, it’s impossible to interpret it and make it more resilient.

That’s how 9’s came to be.

A 99% performance score for a piece of software means it’s not available for 3.7 days in a year.
A 99.9% performance score means it’s not available for 8.7 hours in a year.
A 99.99% performance score means it’s not available for 52.6 minutes in a year. See how by adding 9s you reduce the downtime?

Hence, Last9. Our goal is to add to your 9s to the last 9, so your software is reliable.

Our mission is to make running systems at scale, fun, and ^*embarrassingly easy.

^*Embarrassingly parallel, funnily enough, is how the space we operate in can succeed. :) For that conversation, I’m on X.