Troubleshooting
This page groups common issues by symptom. Start by checking the API key, Base URL, model name, and account balance.
Authentication Failure
Common causes include:
- The API key was copied incompletely.
- The request header is missing
Bearer. - The current terminal did not load the environment variable.
- The token was disabled, deleted, or expired.
Check with:
bash
echo "$DDDT_BASE_URL"
echo "$DDDT_API_KEY"Never send a complete API key to someone else.
Insufficient Balance or Quota
If a request reports insufficient balance, check:
- Account balance.
- Current token quota.
- Whether an unexpected script is making repeated calls.
- Cost changes in recent logs.
Disable a suspicious token first if necessary.
Model Unavailable
This usually means the model name is incorrect, the account lacks permission, or the upstream resource is unavailable.
To resolve it:
- Confirm the model name in the dashboard.
- Switch to a lightweight model to test the connection.
- If you selected a provider prefix, return to the default model name.
- Check the actual error in the logs.
Request Rate-Limited
If you receive a 429 or similar error:
- Reduce concurrency.
- Add a request queue.
- Retry with exponential backoff.
- Check the account tier and token limits.
Request Timeout
A timeout can originate in the client, network, gateway, or upstream model. You can:
- Shorten the prompt.
- Reduce the maximum output length.
- Switch to a faster model.
- Configure a longer timeout for streaming output.
- Record the request time and error response before contacting support.
Unexpected Costs
Check Usage Logs first and confirm:
- Whether a high-cost model was called.
- Whether output token usage is unusually high.
- Whether a test script made repeated calls.
- Whether an unused token was exposed.
If you find unexpected activity, disable the related token before continuing the investigation.