“I know I should apply continuous improvement to operations. But where do I get started?”
To answer this, you can look at the 80/20 rule (80% of your issues are caused by 20% of the causes).
But that begs the question: “What is that 20%?”
It's hard to figure that out because you have a busy team that’s spending most of its time and resources to keep things from catching fire, so they don’t have time for data classification and analysis.
Our new, free product – Incident Insights – solves this for you.
It pulls data from your incident management tool, and in 2 minutes, it applies unsupervised learning to the report.
“What is unsupervised learning?”
It means you don't have to label anything.
The tool reverse engineers the patterns used to create the subjects of your incidents and generates regular expressions to cluster them together.
This gives you a clearer sense of:
- How many things are happening in each cluster?
- How much time is being spent in each cluster?
- Which alarms need to be turned off or have their thresholds adjusted?
- Which incidents need to be routed to a human?
It can also help you see where you need to turn an imprecise alarm like “CPU is high” into a precise alarm, like:
- CPU is high because of the JVM process. This can be automatically fixed.
- CPU is high, and everything is just fine because the response rate is good.
- CPU is high and you don't know what's going on. So you need someone to take a look at it.
That’s how Incident Insights helps you remove noise and increase signal, making your team more productive and reducing costs by decreasing toil.