The product label for this page is Usage. It gives you a record of every tool call your agent has made through ClawLink. Use it to verify what ran, inspect failures, and understand latency.
What logs show
Each log entry represents a single API call. The logs page shows at a glance:
- Total Requests — the total number of calls recorded
- Success Rate — the percentage that succeeded
- Avg Latency — the average response time across all calls
The log table lists individual requests with all the detail you need to understand what happened.
Reading a log entry
Each row in the logs table contains the following fields:
| Field | Description |
|---|
| Integration | The service that was called (for example, gmail, slack, github) |
| Action | The specific operation that ran (for example, send_email, create_issue, list_customers) |
| Status | Whether the request succeeded or failed |
| Latency | How long the request took to complete, in milliseconds |
| Timestamp | When the request was made |
Filtering logs
Use the filter controls to narrow down what you see:
- Search — type any text to filter by integration name or action
- Integration — select a specific service to see only its requests
- Action — filter by a specific operation
- Status — show only successful requests, only errors, or all requests
When you see errors
If a request shows an error status, check the following:
- Open the Connections page and confirm the affected integration is still showing as Connected.
- If the integration needs attention, reconnect it by running the hosted flow again.
- Check that your agent is calling a valid action for that integration.
Use filters to isolate one integration or action. If you see a sudden spike in failures from one service, it usually means the connection needs to be refreshed or the provider had an outage.
