Anomaly Detection
The ODJ provides cost anomaly detection. This ensures that once costs for a specific service in one of your environments rapidly increases, the ODJ makes sure to inform the relevant person(s). Every application created in the ODJ will run in at least 1 environment, usually 2 (WORK, LIFE). As these environments are decoupled from each other, when it comes to costs at service level, the ODJ is differentiating them. This means we show what you spend for a specific service in each of your environments as these costs might differ. Once the costs for a service increases by an amount that exceeds a normal cost trend, the ODJ assumes a cost anomaly and tags the occurrence and the cost increase that is resulting. It’s up to you to decide whether the cost increase was expected or unexpected and requires treatment or other measures.
How it works
We are calculating the rolling average of your daily costs for the past 14 days. With this value, the ODJ will calculate the variance and then derive the standard deviation of your costs.
The Variance is defined as the average of the squared differences from the Mean. The Standard Deviation is defined is a measure of how spread out numbers are.
The ODJ will trigger a cost anomaly, once the daily costs spend for a specific service in one of the environments has reached a so called upper bound. The upper bound is the threshold. Once it is reached, an anomaly will be triggered. The upper bound lays at at least 2 standard deviations above the average daily cost as explained above. In the image below, the upper bound can be referred to as the red lin inside the graph.
Here is an example of How to calculate the upper bound:
Impact of cost anomalies
An anomaly can occur for every service and in every Environment where its consumed, as already explained above. This means that in certain circumstances, there could be multiple open cost anomalies occurring for one of your Products. To make sure that cost anomalies are taken serious and the impact is quickly limited, we have made the total costs that is caused by open anomalies to our UI.
