Younghun

I encountered this error with async api in `fastapi + uvicorn + gunicorn`.(pymilvus version `2.5.8`) @ahmadcolourbox I ask for your code snippet. your function `self.client.search` is sync api not async,...

@ahmadcolourbox really thanks for your detailed guide! I succeed in pymilvus sync + asyncio due to your code snippet. And a milvus developer(thanks for @XuanYang-cn beforehand) is assigned to this...

I got the same issue. I run rest api for unloading all models in triton and request with `v2/health/ready` and `v2/health/live` but both status-code is still 200.

I more investigate and found the confusing moment. At first, I run triton as deployment and attach service and endpoint to the triton. After running well, I enter into the...

@eladamittai Hi. Thanks for my suggested problem.I can't understand your saying. Can you explain to me in detail about your environement?

@jaelgu Yes. I wrongly write my final steps of pipeline. Lastly, I tested it by your suggestion(create indexing earlier than bulk-insert) but the error is still. I revised my issue...

@yanliang567 Thank you for fast analysis! I will test it after following resource increase. > 1. increase cpu of mixcoordinator from `500m` to `6core` > 2. increase cpu of querynode...

@yanliang567 Oh sorry. I add more memory and attach it again [milvus-log.tar.gz](https://github.com/user-attachments/files/24101389/milvus-log.tar.gz)

@yanliang567 I tested it again and attached full logs. In this log, the proxy and mixcoord log is so long. Can I check this tar file? Of course, the latency...

@yanliang567 Above the comment, I tested it on my dev environment that bulk-insert only one collection consisted of 1.8M entities. For the details, I also attach the logs of milvus...