Younghun
@yanliang567 Sorry for late reply. I continuously tested it again after increasing resources but the issue is still.. Can you check my full logs in the attachment? I try to...
@yanliang567 @xiaofan-luan I noticed something odd. The batch pipeline scheduled for early this morning failed. Interestingly, among [the steps I mentioned earlier,](https://github.com/milvus-io/milvus/issues/46284#issue-3718012373) everything succeeded except for step (6): switching the...
@yhmo thanks for your suggested metrics. First, I want to deep dive why only altering alias cause the latency spike. So, I only run altering alias from collection A to...
@yhmo The informations is below. > user_embedding_2025_12_16_2025_12_17_03_43 collection info: {'collection_name': 'user_embedding_2025_12_16_2025_12_17_03_43', 'auto_id': False, 'num_shards': 1, 'description': '', 'fields': [{'field_id': 100, 'name': 'user_id', 'description': '', 'type': , 'params': {'max_length': 100}, 'is_primary':...
@yhmo Right, Most of my application interfaces for query to milvus are all setting embedding field in `output_fields` argument. There are examples of my application interfaces. For refer, you said...
@yhmo Your explanation is perfect to me. I got it! I changed from `disable` to `sync` and test altering alias again. As a result, the dashboard metrics is so awesome....
@haorenfsa Hi, I am ready to migrate from helm-chart based to custom-resource based installation. And I encountered this issue. I read all your discussions and I found this is not...