Logs
You can inspect each one of your worker tasks, see your interaction history, and retrieve past deliverables and responses.
The Logs page gives you a history of executions and all the details related to integrations used to perform a certain task.
General details
For each worker execution you'll see:
Worker name
Identifies the worker
Status (in progress, completed, failed)
Understand if a worker has delivered its task
ID
The unique identifier of the worker. Click on the value to edit your worker.
Run ID
The unique identifier of this specific interaction between you and the worker. Click the email icon to email this worker. Click on the value to talk to the worker.
User ID
The unique identifier of the Toolhouse user who's interacting with the worker (usually you).
Duration
How much time it took the worker to complete the task
Started
The date and time you first requested this task
Last response
The date and time you last heard from this worker
Traces
The Traces section will show all the integrations the worker used to perform the task. You can use this to optimize your integration usage.
When you click on a trace, you can also see what inputs the worker gave to each integration, and you can see the output from the integration itself.
For each trace, you'll see:
Position
The order each integration was called
Integration name
The name of the integration
Intermediate time
Tells you at what time the integration was called. For example, +8min 30s · 20:36:04 indicates that the integration was called at 08:36:04 PM, which is 8 minutes and 30 minutes after the worker started.
Waterfall
Shows you at what point of the overall duration each integration was called, and how long it took
Duration
How long this integration took to give the its answer back to the worker
Input
The data the worker prepared to make this integration work
Output
The full result from the integration
History
The History section contains the full activity from your worker. For autonomous work, you'll see all the steps the worker took; if the task was supervised, you'll see your activity (for example approvals and directions).
User input
Contains your initial request and any other approvals or direction you gave to the worker. For autonomous work, it contains a default start message.
Worker output
The output from your worker. This could be an intermediate result (for example, the worker is thinking, or it's asking you for approval or additional details)
Integrations
Shows what integrations were used by the worker that helped it create its answer
Last updated