> ## Documentation Index
> Fetch the complete documentation index at: https://docs.scitix.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# View Metrics
> Learn how to view aggregated metrics, usage trends, token distribution, and latency performance for ScitiX Model Inference.
After signing in to the console, open the [Metrics](https://console.scitix.ai/model-inference/metrics) page to view aggregated metrics and trend data for model calls.
* **Metrics**: Use this page to observe overall usage, throughput changes, and latency performance over a selected time range.
* **Home**: Use this page for a quick view of major traffic sources, including Top Models, Top API Keys, and member usage. Member usage is visible only to administrators.
* **Usage Detail**: Use this page to inspect individual request records. Go to the [Usage Detail](https://console.scitix.ai/model-inference/usage_detail) page when you need request-level details.
## Filters
You can view metrics by the following dimensions:
* **Time range**: Select the last 24 hours, last 7 days, last 30 days, or define a custom range with a start date and end date.
* **Model**: View all models, or select a specific model to analyze its metrics.
## Summary Metrics
The top of the page shows summary metrics under the current filters:
| Metric | Description | What to Watch |
| ------------ | ------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Total Tokens | Total token usage within the selected time range. | Use this to observe overall usage and cost changes. If total tokens increase sharply, check Token Distribution to see whether the growth comes mainly from input, output, reasoning, or cache-related tokens. |
| API Requests | Total number of API requests within the selected time range. | Use this to understand traffic volume. If requests increase unexpectedly, compare the timing with product releases, batch jobs, or retry logic. |
| Avg RPM | Average requests per minute. | Use this to observe request throughput. If RPM approaches the limit of an API key or service, consider adjusting request pacing or rate limits. |
| Avg TPM | Average tokens per minute. | Use this to observe token throughput. A higher TPM usually means longer context, longer outputs, or more concurrent requests, and may also increase cost. |
## Charts
The Metrics page includes trend and distribution charts that help you analyze calls from different angles:
| Chart | Description | What to Watch |
| -------------------------- | ------------------------------------------------------------------------------------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Token Usage | Shows how token usage changes over time. | Use this to locate usage peaks. If a peak looks abnormal, compare it with application logs and request records in Usage Detail. |
| API Requests | Shows how the number of API requests changes over time. | Use this to understand traffic changes. If requests increase but token usage stays mostly flat, short requests or probe-style requests may have increased. |
| RPM | Shows requests per minute over time. | If RPM rises sharply, request volume or concurrent usage has increased. Compare it with the RPM limits on your API keys. |
| TPM | Shows tokens per minute over time. | If TPM rises sharply, overall token throughput has increased. Check Token Distribution to see whether input, output, or reasoning tokens are driving the change. |
| Token Distribution | Shows the distribution of token types, such as input, cache read, cache write, reasoning, and output tokens. | Use this to analyze cost composition. If output or reasoning tokens take a large share, review generation length, reasoning model usage, and parameter settings. |
| Prompt Bucket Distribution | Groups requests by prompt length bucket and shows the corresponding TTFT and E2E latency percentiles. | If long prompts account for a large share of requests, consider compressing context, splitting tasks, or using Prompt Caching to reduce repeated context processing. |
| TTFT | Shows the time from request start to the first output token. | If TTFT increases, model API callers may notice that responses start more slowly. Analyze this together with model choice, prompt length, streaming behavior, and traffic peaks. |
| E2E Latency | Shows the total time from request start to response completion. | If E2E latency increases, it is often related to output length, model type, context length, or concurrent load. Compare it with output token counts and request records in Usage Detail. |
## Common Analysis Paths
* **Analyze cost changes**: Start with Total Tokens, TPM, and Token Distribution to determine whether the increase comes mainly from input, output, reasoning, or cache-related tokens.
* **Troubleshoot slower responses**: Start with Time to First Token (TTFT) and End-to-End Latency (E2E). When TTFT is high, focus on how quickly the first token returns. When E2E latency is high, focus on output length and full generation time.
* **Monitor traffic fluctuations**: Use API Requests, RPM, and TPM together to identify traffic peaks. If needed, check whether your application has batch jobs, scheduled tasks, or retry amplification.
* **Optimize prompt length**: Check Prompt Length Distribution. If long prompts make up a large share of requests, consider compressing context, splitting tasks, or reusing cached context.
* **Compare model performance**: Filter by model and compare latency, throughput, and token distribution to evaluate whether each model fits your workload.