What to Measure

We can measure all sorts of things in our applications, but which ones are valuable? It depends on your context!

If Linko is your startup's core product, you'll probably want to measure some things for Business Intelligence, such as:

• How many new users are signing up?
• How often do users cancel their account?
• How often are they sharing shortened links on social media?

From an operations standpoint, we're more likely interested in things like:

• How many shortening requests do we get?
• How many unshortening (redirect) requests do we get?
• How many requests fail?
• How long do requests take on average?

There are a few models we can use to focus on which metrics matter most: R.E.D. and U.S.E..

RED – Rate, Errors, Duration

The RED Method applies most closely to services that perform requests, like our Linko service.

• Rate: The number of requests per second.
• Errors: The number of those requests that are failing.
• Duration: How long those requests take.

In this model, Rate and Errors are counters, and Duration is typically represented as a histogram.

USE – Utilization, Saturation, Errors

The related USE Method applies more to finite resources – things like CPU, memory, and disk space.

• Utilization: The percentage of time the resource is busy or the percentage of the resource consumed.
• Saturation: The amount of work the resource has to do, such as queue length.
• Errors: The count of error events.