Skip to content

Request Logs

Overview

The Request Logs page provides a comprehensive view of all API requests made to the Wuliang AI platform. You can filter, search, and analyze individual requests to debug issues, monitor performance, and understand usage patterns. Each log entry contains detailed information about the request, response, and internal processing trace.

Request logs overview

Prerequisites

  • A registered Wuliang AI account
  • At least one API request has been made. See Make API Calls for instructions.
  • Logged in to the console

Access Request Logs

  1. In the console sidebar, navigate to Resources > Request Logs.

    Navigate to request logs

  2. The request logs page loads with your most recent API requests.

    Request logs loaded

Use the filter bar at the top of the page to narrow down the log entries.

Available Filters

FilterDescription
ApplicationFilter by application name
API KeyFilter by a specific API Key
Capability TypeFilter by capability (Text Generation, Image Generation, etc.)
Request IDSearch by exact request ID
Date RangeFilter by time period (today, last 7 days, last 30 days, custom range)
SourceFilter by request source (OpenAPI, Playground, SDK)

Filter bar

Apply Filters

  1. Click a filter dropdown to view available options.

    Click filter dropdown

  2. Select the desired filter value.

    Select filter value

  3. Multiple filters can be applied simultaneously. Active filters are displayed as tags.

    Active filters

  4. Click the Reset button to clear all filters.

    Reset filters

Request List

The request list table displays your API requests with the following columns:

ColumnDescription
Request IDUnique identifier for the request (click to view details)
TimeWhen the request was made
Capability TypeThe type of API capability used
StatusSuccess or failure indicator with color coding
Tokens (Input)Number of input tokens consumed
Tokens (Output)Number of output tokens generated
LatencyTotal response time in milliseconds
Client IPIP address of the requesting client

Request list table

Sort and Paginate

  1. Click column headers to sort the list by that field.

    Sort by column

  2. Use the pagination controls at the bottom to navigate through pages.

    Pagination controls

Request Detail Drawer

Click any row in the request list to open a detail drawer with complete request information.

Click request row

Request Information

The detail drawer displays the following sections:

SectionDescription
Request IDUnique identifier
TimestampExact date and time of the request
ApplicationThe application associated with the request
API KeyThe key used (masked for security)
ModelThe AI model called
Capability TypeThe API capability used
SourceWhere the request originated (OpenAPI, Playground, SDK)
StatusSuccess or failure with HTTP status code
Client IPIP address of the requester

Request detail sections

Request Parameters

The full request body as sent to the API, including model, messages, temperature, and all other parameters.

Request parameters

Response Data

The complete API response including generated content, token usage, and finish reason.

Response data

Error Details

For failed requests, the error code, message, and troubleshooting information are displayed.

Error details

Token Usage Breakdown

A detailed breakdown of token consumption:

MetricDescription
Input TokensTokens in the request prompt
Output TokensTokens in the generated response
Total TokensCombined input and output tokens
CostMonetary cost of the request

Token usage breakdown

Request Trace Timeline

For each request, a trace timeline visualizes the internal processing stages. This is particularly useful for diagnosing performance issues or understanding where failures occur.

Trace timeline overview

Trace Spans

Each span in the timeline represents a processing stage:

Span FieldDescription
NameThe processing stage name (e.g., Authentication, Model Routing, Inference, Response)
DurationTime spent in this stage
StatusSuccess or failure
Error InfoError details if the span failed

Trace span details

Using the Trace for Debugging

  1. Identify slow requests by sorting the list by latency (descending).

    Sort by latency

  2. Open the request detail and examine the trace timeline.

    Examine trace

  3. Identify which processing stage consumed the most time or where the error occurred.

    Identify bottleneck

  4. Use the request ID to correlate with your application logs for end-to-end debugging.

    Request ID correlation

Common Use Cases

Debug a Failed Request

  1. Filter the list by Status to show only failed requests.

    Filter by status

  2. Click the failed request to open the detail drawer.

    Open failed request

  3. Review the error details and trace timeline to identify the root cause.

    Review error

Investigate High Latency

  1. Sort the list by Latency in descending order.

    Sort by latency

  2. Open the top entries and examine the trace timeline to find the slowest processing stage.

    Examine trace

Track Usage by Application

  1. Filter by a specific Application to see all requests for that app.

    Filter by app

  2. Review the token consumption and cost breakdown.

    Review consumption

Notes

  • Request logs are retained for 30 days by default. Contact support for extended retention.
  • The detail drawer masks sensitive information such as full API Keys.
  • Log data may have a slight delay (up to 5 minutes) before appearing in the list.
  • The request ID can be used to search for specific requests across different filter combinations.
  • Trace timeline data is available for all requests, regardless of success or failure.