Built-in Kyvos semantic models for monitoring
This documentation is for Kyvos 2026.8.x. To view documentation for other Kyvos versions, click here. |
|---|
Overview
Kyvos comes with certain built-in semantic models to help you monitor the Kyvos system. These semantic models include all the relevant process details, such as entity information, process type, version, user, status, size, elapsed time, browsable time, memory usage etc. You can view these details in the visualization sheet and share them via email.
To view built-in semantic models for monitoring, perform the following steps.
From the Toolbox, click Semantic Models.
In the Search box, type to find the built-in semantic model.
On this page:
Types of built-in semantic models for monitoring
Process Details
Provides a process details of the semantic model such as build information, entity information, build details, size details and so on.
You can now view skipped build status in Process details semantic model.
Process History
Provides a process history of the semantic model process, including processing information such as process start time, end time, process type, and so on.
Query Analyzer
Provides a historical view of field details, query aggregate details, query details, query engine details, and subquery details. Additionally, it also provides details of the currently running query. This semantic model includes various semantic models through which you can view historical details and currently processing details of the queries.
The Query Analyzer also enables users to analyze the historical execution of natural language (NL) queries across Deep Dive and Precise modes. It provides insights into query execution, token consumption, and performance metrics, helping monitor AI usage and optimize system performance. Using this semantic model, users can:
Analyze the historical execution of natural language (NL) queries for both Deep Dive and Precise modes using worksheets.
Perform cost analysis by monitoring the tokens consumed for NL queries, with the ability to filter the results by AI Space and/or user.
Perform performance analysis by analyzing the execution time of NL queries, with the ability to filter the results by AI Space and/or user.
The Query Analyzer uses the following semantic models:
NL Query Request Details
NL Query Steps Details
NL Query Request Details | NL Query Steps Details |
|---|---|
|
|
Query failure reasons
You can view 'Failure Reason' on the built-in Historical Query Details dashboard. This column shows the details of the reasons leading to query failures.
System Monitor
Monitors the CPU usage, load average, and heap memory usage of the Kyvos cluster's Query Engine and Analytical Server when queries are executed during a specific time period.
System Usage
Monitors the maximum, minimum, and average heap memory used for analytical server and Query Engine.
Note
For the Process History semantic models, you must schedule ‘Process Full Data' for the first time, and later you need to schedule 'Process Incremental Data’.