Can you try clearing Redis cache? this can be cause by lock in the cache that wasnt properly cleared

From the code I notice there is no flow path for what has to happen after the MAX_ATTEMPTS has been reached. I can make a PR that fixes that, since it should probably fall back to attempt fetching the schema itself. In a single container the retry logic doesn't make sense either, since there could be multiple requests waiting on the getSchema, causing all of them to be very close together when timing out.

I just realised that the schema isn't being cached to redis at all. Seems like this await sharedSchemaCache.set('schema', schema); is missing from the setSchemaCache function

Before diving into the code have you tried clearing the redis cache as asked?

Seems like this await sharedSchemaCache.set('schema', schema); is missing from the setSchemaCache function

Thats not missing i think as it's likely not supposed to do that, the actual schema is stored with localSchemaCache where sharedSchemaCache stores a hash to sync multiple instances.

Yes I did clear the cache.

I had this exact same issue when upgrading to 10.10.5 for our development setup. I ended up removing the CACHE_SYSTEM_TTL env which was previously set to CACHE_SYSTEM_TTL=100 and restarted the container, and haven't seen this issue since.

This is related to #22161