Observability

Inferable provides multiple ways to observe and monitor your workflows, making it easier to debug and maintain them in production.

​

Local Logging

You can provide an optional logger instance when creating a workflow to get detailed insights into workflow execution on each machine. The logger should implement the following interface:

Here’s how to configure a logger for your workflow:

When running the same workflow on multiple machines, each instance will log independently on the machine that’s executing that particular part of the workflow. This distributed logging approach helps you track workflow execution across your infrastructure.

​

Web Interface Observability

Each workflow execution creates agent.trigger()s that are visible in the Inferable web interface at:

open https://app.inferable.ai/clusters/$CLUSTER_ID/runs

The web interface provides:

• A list of all workflow runs
• Detailed telemetry for each run
• Agent execution details
• Step-by-step execution flow

​

Coming Soon

We are actively working on enhancing workflow event observability through the Inferable web application. This will include:

• Consolidated workflow event views
• Advanced filtering and search capabilities
• Real-time workflow monitoring
• Enhanced debugging tools

You can track the progress of these improvements in our GitHub repository.