​
Overview

If you have received an email stating that “Endpoint XYZ is running into errors” with a distribution of HTTP status codes, this guide will help you understand what it means and how to address the issue.

​
What does this mean?

The Glean Assistant calls the endpoint mentioned in the email to make requests to your LLM provider (Azure, OpenAI, Google Vertex AI, or Amazon Bedrock). You are receiving the alert because the endpoint has been returning errors. Glean does not have administrative access to your LLM provider, so we need your help to resolve these issues.

​
How to resolve the errors

​
Status Code 401/403

These status codes usually indicate an authentication issue with the endpoint. Please check that your account with your LLM provider is still active and the Admin > Setup > Assistant > LLM Provider settings are correct.

​
Status Code 429

This status code indicates that the endpoint might be underprovisioned for your traffic. Please refer to Glean Assistant LLM Providers to check if you are appropriately provisioned. Contact a Glean representative if you believe you are provisioned correctly but are still receiving 429 errors.

​
Status Code 500

This status code indicates a server error on the LLM provider’s side. Check if there was any outage or service disruption from the provider: Once the errors are resolved, you will stop receiving the alert emails and no further action is required.
Looking for the original version of this page? You can find the archived version here.