Event log gives you a visual summary of all the events occurring during a model interaction session.

There could be multiple events in a session with the ability to see details of a specific event. For example: details of successful connection to foundational models, sending prompts and receiving responses, failed and successful sessions, token usage and other event details. The ability to monitor and view these details enables you to analyze the data and proactively take preventive action for data-security, governance, and compliance.

Event log table details

You can refine your search by available search parameters such as, by Session ID to view all event types specific to the selected session.

Note: Use your cursor to roll over the action icons to identify specific functions for each.

The Event log table sorts and displays details as per these columns:

Event log table details

one

Status: Shows the success or failure of an event.
two

Time: Shows the start time of the event.
three

Event type: Shows each session type such as: Send prompt, Response received, Model connection, and Trigger prompt template.

You can click the Event type link to drill-down and view additional event details.
four

Model name: Shows the foundational model version name that was interacted with, such as: Claude for Amazon Bedrock, VertexAI forGoogle Vertex AI, GPT 3.5 Turbo for Azure OpenAI, and others.
five
Prompt type: Shows the type of prompt used in the automation. These are:
  • Free form prompt logs captured for all generative AI Packages.
  • Prompt Template prompt logs captured for Prompt Templates.
six

Automation name: Shows the automation name that used a specific prompt.
seven

Session ID: Shows the automation session ID.
eight

Device name: Shows the ID of the device on which the automation was executed.
nine

User: Shows the ID of the user who ran the automation.
ten

Folder path: The folder location of the automation from where it was executed.

Supported event type logs

The following event type logs are captured for all model interactions in an automation.
  • Model connection
  • Send prompt
  • Response received
  • Trigger Prompt Template

In the Event type column, click the link to drill-down further to view event details by the following parameters. The details vary as per event type. On the same screen, view different parameters for each event type.

Event log details

Note: The Automation type field shows the type of automation used in the model interaction session. The remaining fields are same as the Event log table details.

Free form logs (generated when using Generative AI Packages) are captured for automations executed with generative AI Packages, and is supported by Model connection, Send prompt, and Response received event types.

Each event displays the following:
  • Status: Indicates whether the model connection was successfully established.
  • Event type: Categorizes the event as a Model connection.
  • Time: Specifies the timestamp of the connection attempt.
  • Prompt type: Identifies the type of prompt used.
  • Model name: Specifies the name of the language model used. For example: GPT-4.
  • Session ID: A unique identifier for the connection session.
  • Automation Name Indicates the name of the automation.
  • Device Name: Specifies the device associated with the automation.
  • User: Identifies the user who initiated the automation.
  • Automation Type: Categorizes the automation type. Such as Task Bot, API Task.

Let us go over the supported event type parameters.

Model connection
  • Publisher: Shows the foundational model publisher name, such as Amazon Bedrock, Google Vertex AI, Azure OpenAI, OpenAI and others.
  • Connection method: Shows the connection type used to connect to a model. For example: Amazon Bedrock, Azure OpenAI, and OpenAI connect using API keys. Google Vertex AI connects with a User Account, Service Account or Control room OAuth connection.
Send prompt
  • Prompt: Shows the complete prompt-text entered by the user.
  • Click More details to view the following information:
    • Publisher: Shows the foundational model publisher name, such as Amazon Bedrock, Google Vertex AI, Azure OpenAI, OpenAI and others.
    • Request configuration: Shows the parameter values used for the prompt, such as: Max tokens, Top P, Temperature and others. These parameters vary for different models.
Response received
  • Prompt: Shows the complete prompt-text entered by the user.
  • Response: Shows the response received from the model.
  • Click More details to view additional information about Tokens consumed: Shows the number of tokens consumed by the model interaction. This helps keep track of tokens consumed and balance.
Trigger AI Skill
  • Trigger AI Skill details: Shows the details of the prompt used when testing the model connection.