A user is reporting that their change stream is failing with a ReadConcernMajorityNotAvailableYet errors:

We have a service that listen on one collection, we use change stream to watch for a collection. We use mongodb 5.0 version. I don’t know why but we encountered this error:

pymongo.errors.OperationFailure: Executor error during getMore :: caused by :: Committed view disappeared while running operation, full error: {'ok': 0.0, 'errmsg': 'Executor error during getMore :: caused by :: Committed view disappeared while running operation', 'code': 134, 'codeName': 'ReadConcernMajorityNotAvailableYet', '$clusterTime': {'clusterTime': Timestamp(1635457764, 1), 'signature': {'hash': b'\xb9\xedq\xf5)\xa2\xd9s\xe6\xe8\xbc\x01\xfd\x1b\xb3\x1d[;#V', 'keyId': 7021065227265376259}}, 'operationTime': Timestamp(1635457764, 1)}
About database server we use:

storage engine: WiredTiger
Replication Factor: Three-node
Minor Version 6.0.0-20210924172808_1
Version 5.0
Read-only Nodes: 0

I would like to get help on knowing what could be the cause of the error and how to prevent it?

https://www.mongodb.com/community/forums/t/committed-view-disappeared-read-concern-majority-not-available-yet-code-134/130471

Should the server be attaching the ResumableChangeStreamError label to ReadConcernMajorityNotAvailableYet errors so that the driver can retry?