Skip to content

Handle when the embedding service is out of GPUs #139

Description

@Jonathan-Adly

When the embedding service is out of GPUs. We get an error message:

['Failed to save pages: ["Failed to get embeddings from the embeddings service. Repsonse: {'id': 'sync-4bb0e1d2-d4b4-4ed8-9236-40826f9542d8-u1', 'status': 'IN_QUEUE'}"]']

We want to handle this via webhooks: https://docs.runpod.io/serverless/endpoints/send-requests. Probably will be simpler if we handle all ingestion requests to the embeddings service via webhooks

{
  "input": {},
  "webhook": "https://URL.TO.YOUR.WEBHOOK"
}

This is a rare bug, so priority is low.

Metadata

Metadata

Assignees

Labels

bugSomething isn't working

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions