Metrics
The Usage page on the NextAI platform provides detailed insights into your usage of our services, helping you understand and optimize your AI model deployments. This document outlines the types of metrics available and how to interpret them for better resource management and planning.
Overview
NextAI’s Usage page is designed to offer comprehensive visibility into your interactions with our platform, focusing on three primary areas:
- User Usage Metrics
- API Usage
- Tokens Usage
- Deployed Models Metrics
Each section provides valuable data over a specified date range, allowing for a granular analysis of your activities.
User Usage Metrics
1. Date Range Selection
1. Date Range Selection
At the top of the Metrics page, you can specify the date range for the data you wish to view. This feature enables you to focus on specific intervals for detailed usage analysis. For instance, you might select from January 1, 2024, to February 22, 2024, to examine your activities within that period.
2. API Usage
2. API Usage
This section tracks the total number of API calls made to NextAI services. API calls are any requests sent to NextAI for model inference, data retrieval, or other interactions that utilize our APIs.
Total API Calls
Definition: The sum of all API requests made during the selected date range.
Importance: Helps in understanding the volume of interactions and potential scaling needs.
3. Tokens Usage
3. Tokens Usage
In the context of AI models, tokens represent the units of text processed or generated by the models. This section breaks down token usage into two categories:
Total Tokens
Definition: The cumulative count of tokens processed, including both prompt and completion tokens.
Importance: Provides insight into the extent of data processed by the AI models, which can influence cost and resource allocation.
Prompt Tokens
Definition: The number of tokens used in the prompts sent to the AI models.
Importance: Reflects the complexity and size of the inputs provided to the models.
Completion Tokens
Definition: The number of tokens generated as outputs by the AI models in response to the prompts.
Importance: Indicates the volume of data produced by the models, useful for understanding output data handling and storage needs.
Deployed Models Metrics
This area offers insights into the usage and performance of your deployed AI models within NextAI. Before accessing the metrics, you need to select the specific cluster where your models are deployed.
Cluster Selection
Since models can be deployed across different clusters for scalability and redundancy, selecting a cluster allows for focused analysis of model performance and resource utilization in that specific environment. Once a cluster is selected, you can view various metrics related to the models deployed in that cluster, such as the number of inference requests, processing times, and resource consumption.
Conclusion
By regularly monitoring these metrics, you can gain insights into your usage patterns, identify potential bottlenecks, and make informed decisions about scaling, optimizing, and budgeting for your AI projects. If you have any questions or need further assistance with interpreting your metrics, our support team is here to help.