As the data dashboard for globally monitoring the running status and usage of agents, the Usage Overview module in Agent Supervision helps administrators grasp the platform's overall operation status and provides data support for agent optimization, resource expansion, and cost control.
You can filter data by multiple time dimensions to flexibly analyze usage trends across different periods.
You can rank agents by usage times and token consumption to identify high-value/high-consumption agents.
You can globally monitor the agent quantity, usage times, and user quantity in real time.
The following figure shows the effect of the Usage Overview page.
Global Overview is used to quickly view the overall running data of agents on the platform, providing an intuitive grasp of the platform's agent usage scale and popularity.
Currently, the following three types of global metric data can be viewed:
Agent quantity: Total number of agents currently created on the platform, reflecting the platform business scenario coverage.
Agent usage times: Total number of interactions between all agents and users within the statistical period, reflecting the platform usage activeness.
Agent user quantity: Number of unique users who have used agents within the statistical period, reflecting the user coverage.
Input/Output token: Total number of input and output tokens used by agents across the platforms within the statistical period.
Click the time filter drop-down box in the upper right corner to filter data by one of the following five time dimensions:
Current Day: View real-time data of the current day.
Last 7 Days: View cumulative data within the last 7 days.
Last 30 Days (default option): View cumulative data within the last 30 days.
All Periods: View all data since the platform's launch.
Custom: Manually select start and end dates to flexibly analyze data in a specified period.
Top five gents are displayed by usage times in the descending order, helping identify high-value business scenarios.
(1) Filter dimension: You can select Current Day, Last 7 Days, Last 30 Days, All Periods, or Custom.
(2) Leaderboard content: Agent names and their corresponding usage times are displayed by ranking.
(3) Application scenarios:
Identify the frequently-used agents to prioritize their performance and resource allocation.
Analyze agents with low usage times to optimize their functions or evaluate whether to turn them offline.
Combined with business activities, analyze the reasons for fluctuations in usage times.