> ## Documentation Index
> Fetch the complete documentation index at: https://docs.aihubmix.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Request logs and troubleshooting

> Find calls by time, model, key, and status. Review tokens, cost, and latency, configure the table, export records, and share a Tid with support.

Request logs record model-call outcomes and usage. Start with the relevant request when investigating failures, latency, or charges.

<h2 id="find-request">
  Find a request
</h2>

1. Confirm the account and workspace, then open **Request logs**.
2. Select dates and available filters such as model, key, or status, or search for a request identifier.
3. Open a record to review its status, cost, model, and timing.
4. Copy its **Tid** when you need help. Include the time and symptoms in a [support ticket](/en/FAQs/Feedback).

<h2 id="request-fields">
  Read request information
</h2>

| Information | Meaning |
| - | - |
| Status / finish reason | Request outcome and how generation ended; missing values appear as `-` |
| Input / cache / output / reasoning tokens | Usage breakdown returned in the record |
| Cost | Recorded request cost, interpreted with the model's pricing units |
| Total duration / TTFT / throughput | Overall time, wait for the first token, and output speed |
| Member / API key | Request attribution; historical information can identify deleted resources |

A missing field does not mean zero cost, zero latency, or successful completion. Access to log metadata does not imply access to full prompts, responses, or another member's files.

<h2 id="table-and-export">
  Table and export
</h2>

Use column settings to choose visible fields and their order. Supported columns cycle through ascending, descending, and default ordering while keeping filters.

Scroll to load more records. If loading more fails, retry without losing the rows already shown. Check the current scope and filters before exporting; a partial query should not be treated as the entire account's bill.

<h2 id="troubleshooting">
  Common investigations
</h2>

* **Requests fail**: Check key status, model scope, balance, and limits, then consult [HTTP status codes](/en/FAQs/HTTP-Codes).
* **Responses are slow**: Separate time to first token from subsequent output speed and compare similar inputs.
* **A charge is unclear**: Check the actual model, pricing unit, input/output usage, and other billable items. The last displayed result alone may not explain every charge.
* **A record is missing**: Widen the date range, remove extra filters, and check the account and workspace.

Related: [Usage analytics](/en/console/usage) · [API key management](/en/console/api-keys) · [Audit logs](/en/enterprise/audit)

Last updated: 2026-10-13


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.