Articles in this section

Troubleshoot Celigo AI errors

Gogula R Bhavana
Updated

Error code

Error message(s)

Likely causes

Resolution

400 Bad Request

  • Input exceeds the maximum token limit for this model

  • Required body param missing

  • Invalid editor_type, processor_type, or resource

  • Invalid mappings, sourceList, or filter record

  • Exceeded maximum request size. The request body is too large.

  • The input exceeds the model’s maximum token limit.

  • One or more required parameters are missing or invalid.

  • The requested body size exceeds the model’s allowable limit.

  • The request payload is malformed or improperly formatted.

Ensure that your input (the prompt and the expected output) remains within the model’s maximum token and payload limits. If you're working with large inputs, break them into smaller, sequential requests. Additionally, confirm that all required fields and values are correctly filled in before sending the request.

401 Unauthorized

  • Invalid auth

The authentication token or API key is missing or invalid. Ensure you have provided the correct credentials.

Verify your API credentials to ensure they are correct and included in the request header. If necessary, re-authenticate and confirm that your token or API key has the appropriate permissions for the request.

403 Forbidden

  • Unauthorized Celigo token

  • Celigo AI is not enabled

  • OpenAI request rejected by content filter

  • The token is invalid or doesn’t have the necessary permissions.

  • Celigo AI features are not enabled for your account.

  • OpenAI’s content moderation system blocked the request.

Ensure you're using a valid Celigo token with the required permissions. If AI features are not enabled for your account, reach out to your administrator or Celigo Support to request access. Additionally, review your prompt content to ensure it aligns with OpenAI’s safety guidelines.

404 Not Found

  • User not found

  • No relevant tables or operations were found

  • Resource with ID {} not found

The specified resource (such as user, table, operation, or ID) was not found or does not exist.

Make sure the resource ID, name, or reference is correct. Confirm that the resource exists in the intended environment and is accessible through the API.

413 Payload Too Large

  • The request body is too large. Reduce the size of your input

  • Exceeded maximum request size

The request body exceeds the maximum size limit allowed by the API.

If you're sending a large payload, consider reducing the data volume by splitting it into smaller requests that are within the API's size limits.

422 Unprocessable Entity

  • Invalid or malformed JSON

  • Invalid prompt or follow-up prompt

  • Unexpected error while retrieving metadata

  • Could not retrieve connection details

  • Connection offline

  • The model returned JSON that could not be parsed.

  • The prompt is invalid or not properly structured.

  • Backend errors occurred due to connection or metadata issues.

Ensure the JSON structure is correctly formatted and the prompt is well-structured. Also, confirm that all necessary metadata and connections are available. If needed, reconnect or refresh your environment and try again.

429 Too Many Requests

  • The rate limit was exceeded. Wait a few minutes and try again.

  • Too many requests in a short period. Reduce the frequency of your requests.

API request rate limits exceeded (minutes or per day).

If you’ve hit a rate limit, wait a few moments before retrying. To avoid repeated issues, reduce request frequency or use exponential backoff and retry logic.

503 Service Unavailable

  • OpenAI service temporarily down

  • Error fetching Pinecone embeddings

  • The service is currently experiencing an outage or high traffic.

  • A required backend dependency (e.g., Pinecone) is unavailable.

In case of a temporary outage or overload, wait a few minutes and try the request again. Monitor the service status, and consider setting up automatic retries with a delay to handle intermittent issues.

504 Gateway Timeout

  • The request timed out. Wait a few minutes and try again.

  • The request timed out due to extended processing time.

  • The server or network is currently overloaded.

Simplify the input or split it into smaller parts that are easier to process, then wait a few minutes and try again.