Workflow agent logs
View a log of a workflow agent's execution in Mark (Optimizely Agent Platform) to debug the agent and see how it is performing.
Use the Agent activity logs tab to see execution activity across all agent types.
Access your agents
- Go to home.optimizely.com.
- Select your organization.
- Click Mark.
- Click Agents.
- Select the Your Agents tab.
View the logs
Workflow agent logs are accessible from two locations, the Agent Activity Logs tab and the workflow editor.
From the Agents page
To access the logs of a workflow agent from the Agents page, complete the following:
- Complete the steps in the Access your agents section.
- Select the Agent Activity Logs tab.
- (Optional) Use the Agent Type drop-down list to filter by Workflow.
- Click a workflow's execution to view its logs.
See Agent activity logs for filters and column configuration.
From the workflow editor
- Complete the steps in the Access your agents section.
- Click the agent name, or click More > Edit Agent.
- Click More > View Logs.
- Click a workflow's execution to view its logs.
If a workflow agent has an impact estimate, the list of executions includes an Impact column, and the details panel includes an Impact row. See Agent impact reporting.
Export the execution logs
Export a workflow agent's execution logs to a file to keep a record of agent activity, such as tracking performance, validating outputs, or meeting compliance requirements. You can filter the export by date range and choose which columns to include.
- Complete the steps in the Access your agents section.
- Click the agent name, or click More > Edit Agent.
- Click More > Export logs.
- Select a date range:
- Last 7 days
- Last 30 days
- Custom range – Select a From and To date.
A banner in the dialog displays how many executions the selected range includes.
- Select the columns to include in the export. Click Deselect all or Select all to toggle every column at once.
- Select a format from the Format drop-down list, either Excel or CSV.
- Click Export.
The following columns are available for export:
Execution info
- Workflow Name – The name of the workflow agent.
- Execution ID – The unique ID of the execution. Use this ID to identify a specific agent run.
- Execution Time – The timestamp of when the execution started.
- Triggered By – The name of the person who ran the workflow agent. If a webhook, schedule, or email trigger starts the workflow agent, the agent's owner displays instead. See Workflow agent triggers.
- Trigger Type – The type of trigger that started the execution, such as Chat or Webhook.
- Version – The version of the workflow agent that ran.
Input and output
- Trigger Payload – The input data the trigger passed to the workflow agent.
- Output – Mark's response, including any error messages.
- Artifact Links – Links to any artifacts the execution produced.
Performance
- Status – Whether the execution completed or failed. See Execution statuses.
- Duration – Total time the execution took to run.
- Credits – Approximate credit usage for the execution. See Mark credits.
Impact
Select the Impact checkbox to add the following columns. See Agent impact reporting.
- Impact – The impact recorded for the run. In Excel exports, this column holds numbers that you can sum.
- Impact Unit – The unit of measurement for the impact.
- Impact Category – The category of the impact.
Workflow agent log overview
Running
The workflow agent's log that is still running displays similar to the following:
- Return – Go back to the Agent Activity Logs page.
- Name – The name of the workflow agent.
- Status – Indicates the workflow agent is still running. See Execution statuses.
- Timestamp – The timestamp of when the workflow agent started.
- Status – Indicates the workflow agent is still running.
- Cancel Execution – Stop the workflow agent.
- Agent Name – The name of the workflow agent.
- Execution ID – The unique ID of the workflow agent's execution. Use this ID to identify a specific agent run.
- Execution Time – The timestamp of when the workflow agent executed started.
- Triggered By – The name of the person who ran the workflow agent. If a webhook, schedule, or email trigger starts the workflow agent, the agent's owner displays instead.
- Input Variables – The variables that were used in this execution. Click Copy to copy the variables to your clipboard to easily reference and debug agent runs.
- Output – Mark's response including any error messages.
- Execution Memory – The memory chunks the agent transversed during the agent workflow's execution.
- Close details – Close the details panel of the execution.
- Trigger – The trigger that started the workflow agent.
- Agents run in this workflow – The overview of the agents invoked in this execution. They display as gray because they have not run yet.
- Overview map – A smaller, "picture-in-picture" view of the entire workflow agent.
Completed and failed
The workflow agent's log with a Completed or Failed status displays similar to the following:
- Return – Go back to the Agent Activity Logs page.
- Name – The name of the workflow agent.
- Status – Whether the workflow agent completed or failed. See Execution statuses.
- Timestamp – The timestamp of when the workflow agent started.
- Status – Whether the workflow agent completed or failed.
- Agent Name – The name of the workflow agent.
- Execution ID – The unique ID of the workflow agent's execution. Use this ID to identify a specific agent run.
- Execution Time – The timestamp of when the workflow agent started.
- Triggered By – The name of the person who ran the workflow agent. If a webhook, schedule, or email trigger starts the workflow agent, the agent's owner displays instead.
- Duration – Total time of how long the workflow agent took to execute.
- Credits – Approximate credit usage for this workflow. This total accounts for both the workflow orchestrator and specialized agents used. Final usage may vary slightly. See Mark credits.
- Input Variables – The variables that were used in this execution. Click Copy to copy the variables to your clipboard to easily reference and debug agent runs.
- Output – Mark's response including any error messages.
- Execution Memory – The memory chunks the agent traversed during the agent workflow's execution.
- Trigger – The trigger that started the workflow agent.
- Close details – Close the details panel of the execution.
- Agents run in this workflow – The overview of the agents invoked in this execution. Click on agents to see their Execution and Agent Details. Click View Agent to view or modify the agent's details. See Specialized agent logs overview for information.
- Close – Return to the list of the workflow agent's executions.
- Overview map – A smaller, "picture-in-picture" view of the entire workflow agent.
Additional elements
Condition or Loop steps
If your workflow agent includes Condition or Loop logic elements, Mark tracks the following execution statuses in the logs:
- Pending – Not yet started.
- Running – Currently executing.
- Completed – Finished successfully.
- Failed – Encountered an error.
- Skipped – Intentionally skipped.
- Canceled – Canceled by user.
Code steps
Click a code step in the workflow view to open its details panel. The panel reports the following:
- Exit code – The code the sandbox returned. The panel displays exit 0 in green and any non-zero code in red.
- Duration – The execution time in seconds.
- Input Parameters – The resolved value of each variable. The badge reads static for Fixed mode and predicted for Automatic mode.
- Summary – The summary value the handler returned, when it returned one.
- stdout and stderr – The captured output of the run.
- Return Value – The value the handler returned, as formatted JSON.
The Input Parameters section lists every value the orchestrator resolved, including values it then dropped. A value that matches no declared variable displays here but never reaches the handler.
A failing code step produces several entries. Mark retries the step up to three times, and the execution log lists each attempt separately.
For the handler contract, the sandbox limits, and what each error banner means, go to Workflow agent code steps.
Tools
If agents in the workflow include tool calls, the parameter generation logs display within the Thinking chunks of the workflow execution log. The logs capture how Mark interprets input, isolates the required parameters, and formulates the JSON structure passed from one agent to the next. Use this information to verify that parameters were predicted as expected or to diagnose issues when a workflow produces unexpected results.
For example in the following image, Mark distills a user's starting input into a structured parameters object, such as a news_topics field populated with relevant values.
Ask Mark about an execution
Use Mark chat to ask questions about a workflow agent execution in natural language. Chat helps you understand what happened during a run without reading through every step and memory chunk manually.
Start a chat-based debugging session using one of the following methods:
- Open Mark chat from the execution log details view to load the execution context automatically.
- Open Mark chat directly and reference the execution ID in your prompt (for example, Debug execution abc-123).
After Mark loads the context, ask questions such as the following:
- Why did the agent fail at step 3?
- Which workflow step took the longest?
- What were the inputs to the step that failed?
- What was the total token usage?
- Show me the full tool call sequence.
Mark restricts execution debugging to your instance.
Note
If you use Opti ID, administrators can turn off generative AI in the Opti ID Admin Center. See Turn generative AI off across Optimizely applications.