Event logs
Review conversation logs or workspace actions using Audit or Logs in NLX's Dialog Studio
Last updated
Review conversation logs or workspace actions using Audit or Logs in NLX's Dialog Studio
Last updated
The Audit page provides a complete record of the write/delete events that have occurred within Dialog Studio. Having a record of this activity can assist in detecting security violations, performance problems, flaws in applications, and more.
To access the Audit page, click on your username at the bottom of your workspace menu and select Audit from the profile options:
The Audit page lists the following information:
Event: A brief description of the event change in the workspace
Event source: Source reporting the change
Event type: Whether the event was a Write or Delete action
Time + User: Logged time of the event and the user (hover over initial to view user's email)
Apply filters for Start and End times, Principal email, Event type, and Event name to further narrow your search.
Logs provides a detailed view of the conversations occurring with bots in your workspace. While History tracks significant moments of the NLP (detecting user messages and routing to the appropriate intent flow), Logs displays all events processed by the NLX Natural Language Understanding (NLU) technology, including selections a user has made during a conversation, nodes that were traversed, and webhooks/APIs invoked.
When viewing logged events, the results may hold an overwhelming amount of data. Use the filters to auto-update the results to only the information you would like to see:
Event type: Select from the associated dropdown to filter by Event type
Time: Defines the time interval you would like to view, Relative or Absolute. You can switch to the other option by clicking on the ellipsis icon directly above the field. To choose a Relative timeframe, choose an interval from the dropdown field below. To indicate an Absolute timeframe, select the exact Start time and End time from the calendar buttons that are provided
Bot: Filter your view to any single bot within your workspace by selecting from the bot dropdown. To view logs across all bots, you can click the “x” on the right of the field to clear the bot filter
Region: Gives you the option of narrowing your view geographically. Select EU from the dropdown to view only logs based on data gathered within the European Union. Otherwise, leave the default selection at Global to include all logs. You can click on the “x” on the right of the field to clear the filter
Conversation ID, Correlation ID, and User ID: You can enter ID numbers in the text fields to filter for Conversation, Correlation, and/or User
Once you've made your selections, the filtered results will automatically populate in the results.
Click any Event listed within the results to view more advanced details in the modal window that appears from the right:
| EVENT TYPE | DESCRIPTION |
|---|---|
Action invoked | Initiate a webhook call to an action and results in success or failure response |
Action responded | Response to Action invoked call and results in success or failure response |
Auto escalation state | Indicates that escalation was triggered in the conversation tree when it hits a default escalation intent |
Choice auto traversed | Triggers when there is only one available user choice. instead of prompting the user, the application automatically selects the only choice and moves on with the conversation. You may disable this setting on the User choice node. |
Choice selected | User selected from a slot or data request option within the conversation |
Condition evaluated | Records conditions in the conversation tree and results in pass/fail & appropriate values |
Context attributes modified | Indicates change made to context attributes |
Conversation ended | Triggers upon last message sent to user by bot |
Conversation escalated | Indicates that an escalation node was hit in the conversation tree |
Conversation started | Triggers upon first message received from the user |
Default intent retrieval | Indicates that the default intent was triggered |
External processing completed | Corresponding external event has been completed |
External processing hook invoked | Initiate call for the external event to take place; success/failure |
External processing hook responded | Result of the webhook invoked for processing the external event; you can set context using the response |
External processing status updated | Sends status of conversation through webhook to external event |
Fallback state | Occurs when bot cannot determine how to traverse the conversation flow |
Frustration state | Indicates the NLP has identified that the user sentiment is low |
Incomprehension state | NLP did not understand what the user said |
Intent capture state | Next message to the bot is expected to be an intent |
Intent redirection | Conversation flow routes to a different intent |
Lifecycle hook invoked | Webhook that is initiated upon defined lifecycle events: Conversation start, Conversation end, Escalation, Stream state modifications |
Lifecycle hook responsed | Response of the lifecycle webhook call |
NLP invoked | Request is sent to the NLP |
NLP responsed | Response from the NLP |
NLU request received | A message from the user to the NLU |
NLU responded | The response to a user request to the NLU |
Node met all conditions | During node traversal, this indicates the node conditions passed and traversal will occur |
Node traversal | Indicates what node is currently being visited |
Repeat state | Indicates the user has repeated themselves |
Spell auto correction | Indicates the bot has auto-correct enabled and records the result of the auto-correction logic |
State created | Indicates the conversation state has been created. Occurs during conversation initialization |
State retrieved from cache | Indicates the conversation state existed in cache. Occurs when conversation had already been started by a previous message |
State shared with voicecompass | Indicates bot is connected to a voice compass journey assistant and context variables have been shared with the journey assistant |
State values ejected | Indicates values are being cleared from the conversation state. Includes names of values being ejected |
Subtree traversal | Indicates what page is currently being visited in an intent flow |
Variable resolution failed | Triggers if there are unresolved variables or if an error message occurs due to an invalid webhook or technical error |
Variables requested | Initiate a data request webhook call to a variable. Expected response is an object |
Variables returned | Records what has been resolved as part of a data request webhook call |
Wait state | For Amazon Connect, indicates the bot is waiting for input |
Webhook response validation failed | Occurs when there is an invalid payload from the webhook and the handling is unknown |
