Record the average for service ops metrics for each moment in a weekly cycle.

Some metrics, notably operation rates, are periodic. In other words, every monday morning has roughly the same metric, and each Saturday afternoon follows roughly the same pattern.

We can use this data to make better predictions about what is normal and what is an anomaly.

Before

Using 2 sigma:

image

After

Using 1.5 sigma

image

Merge request reports

Loading