> ## Documentation Index
> Fetch the complete documentation index at: https://docs.workflowmachine.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Monitor usage through runs

> Use run history and usage information together to understand how often workflows execute and where credits are being consumed.

## Runs are part of usage awareness

Usage numbers are more helpful when you connect them to actual workflow behavior.

Run history helps you understand:

* which workflows are running often
* whether those runs are producing value
* where unexpected execution volume might be coming from

This is why usage monitoring is not only a billing topic. It is also an operations topic.

## What to watch for

When reviewing runs for usage patterns, look for workflows that:

* trigger more often than expected
* execute too many unnecessary steps
* rely heavily on AI where a simpler approach would work
* fail repeatedly and keep consuming resources

These patterns often point to workflows that should be simplified or tightened.

## Connect run history to workflow design

If a workflow is consuming more run credits than expected, ask:

* is the trigger firing too often
* does the workflow contain unnecessary steps
* is the branching logic doing too much extra work

If AI usage is higher than expected, ask:

* is the AI step essential
* is it running more often than intended
* could the task be narrowed to reduce waste

## Review high-traffic workflows first

You do not need to audit every workflow equally.

Start with:

* workflows that run frequently
* workflows that include AI steps
* workflows tied to production operations

These tend to have the biggest impact on both usage and user trust.

## Healthy usage habits

A few habits go a long way:

* review the first runs after publishing
* revisit workflows that fire on schedules
* simplify workflows that do more work than needed
* fix repeated failures quickly

The goal is not to minimize usage at all costs. The goal is to make sure usage reflects useful automation, not accidental complexity.