Milvus Cluster - Elasticity/Redundant #4136
-
Hello, I have been using haystack and Milvus for vector search. Now I am about to deploy (AWS) what my team developed and I thought I could ask you guys about relevant points to deploy a Milvus server with elasticity and redundancy, so It can serve thousand requests every minute. So what would be a neat advice path to follow. Thanks so much. |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
As you need redundancy, the option will be a Milvus Cluster. It will be elastic, but currently it will only provide manual scaling. But it scales in and out. Milvus team provide a nice calculator tool, where you can do estimates on the hardware requirements based on your project. IMHO, I would go with an EKS cluster running Milvus. But you can run on AWS EC2 instances too. Start small (fitting your immediate need), and scale out the cluster and Milvus as you need. |
Beta Was this translation helpful? Give feedback.
Hi @wilsonlimaneto
As you need redundancy, the option will be a Milvus Cluster. It will be elastic, but currently it will only provide manual scaling. But it scales in and out.
Milvus team provide a nice calculator tool, where you can do estimates on the hardware requirements based on your project.
IMHO, I would go with an EKS cluster running Milvus. But you can run on AWS EC2 instances too.
Start small (fitting your immediate need), and scale out the cluster and Milvus as you need.