Welcome to Tim's documentation

Tim, the Taloflow Instance Manager, is an AI that helps you reduce your AWS costs.

Get Started    


This page describes the monitoring capabilities of Tim. Specifically, calculating Real-time Cost and Engagement Score.

Real-time Cost

In the major components described in the Overview, you may have seen the following:

  • Current cost per hour
  • Live cost graph
  • Most costly tags table
  • Least engaged resources table

Tim uses its own proprietary embedded cost model to calculate real-time cost of your AWS resource every minute. This is analogous to a "live" or "instantaneous" cost, which by default is presented as a cost-per-hour metric for the selected resources.

Generally speaking, the embedded cost model factors in various inputs, including but not limited to:

  • The Type of resources
  • The Resource consumption rate
  • Charging patterns for those Resources from AWS

Engagement Score

The Engagement Score serves the purpose of identifying optimization opportunities faster than any other method available.

In the Overview, you may have seen that the "Least engaged resources" table gives you a quick overview engagement sorted by their cost-per-hour. Less engaged resources are usually low-hanging fruit that can be acted upon.

Acting on "Less engaged resources"

Clicking “see all” in the Overview page will take you to the Resources page where the expanded view of the table will be sorted by the Engagement Score.

How the Engagement Score is calculated

The Engagement Score is based upon rules in Taloflow's rules engine followed by a feedback mechanism that makes the rules more accurate over time. At first, Tim will look at standard usage metrics like CPU, size, cost, disk reads and writes, types of workloads (determining whether they are critical or trivial) and network traffic. As Tim gets more data from your AWS account, it will be able to infer more. For example, if the order processor service hasn't placed an order in five weeks, perhaps that resource is not engaged.