Feature Request: Retry Mechanism for Connector for Temporary Failures

[ChrisBalkIW]

We're using Onyx to index a Bookstack library, and it's working great in general. However, we've noticed a recurring issue when Bookstack goes temporarily offline—for example, during updates or restarts. This is part of my own doing as the (soon to go
away) hypervisor that hosts Docker which runs the Bookstack container takes a very long time to update. When this happens, the Onyx connector attempts to connect once, fails, and then the entire indexing task stops with an unrecoverable error. I don't see any sort of retry or backoff logic to handle these temporary outages gracefully.

Traceback errors:
onyx.connectors.bookstack.client.BookStackClientRequestFailedError: BookStack Client request failed with status 503: Service Unavailable

onyx.connectors.exceptions.ConnectorValidationError: Unexpected BookStack error (status=503): BookStack Client request failed with status 503: Service Unavailable

onyx.background.indexing.job_client.SimpleJobException: Indexing task failed: attempt=433 tenant=public cc_pair=3 search_settings=2

Let me know if you'd like any testing help or further context—I’d be happy to assist.