Kubernetes Anomaly Detection
Updated over a week ago

Finout provides real-time visibility into your costs through your MegaBill and also utilizes your historical data to identify any cost anomalies that may occur. Using machine learning models, Finout automatically detects both increases and decreases in your costs, allowing you to quickly investigate any deviations from your regular spending. You can view these anomalies either in Finout or directly in your Slack channel.

Finout automatically scans your most popular tags and services for all your cost centers and virtual tags. For every newly created virtual tag, an anomaly scan is added, providing full coverage of your data.

Specifically for K8s, The following Kubernetes anomalies are tracked automatically:

  • Deployment

  • Demonset

  • K8s_namespace

  • Cronjob

Pod labels are not tracked automatically; you must request specific labels to be monitored.

How to Review Anomalies

An anomaly can be a result of an unexpected increase or a decrease in cost. You can drill down and see the MegaBill for the costs associated with a specific anomaly. Anomalies can be viewed in Finout or posted as messages to Slack (see How to Configure Sending Anomalies to Slack).

  1. Select Anomalies.

  2. Search for Kubernetes anomalies using the search bar.

  3. To view the MegaBill associated with the deviation, click Investigate.

  4. To delete an anomaly, click the trash can and then click Yes.

  5. To leave a comment on an anomaly, click the comment icon, enter a comment, and then click Save.

How to Configure Sending Anomalies to Slack

Anomalies can be posted as messages to Slack.

To enable anomaly messages:

  1. Create a Slack endpoint (see Create an Endpoint).

  2. Select Use this Endpoint for sending alert notifications.

To learn more about Anomaly Detection, please refer to our main documentation on this topic.

Still need help? Please feel free to reach out to our team at [email protected].

Did this answer your question?