Skip to main content
Version: User Guides (Cloud)

FAQ: Cluster

This topic lists the possible issues that you may encounter while you use Zilliz Cloud clusters and the corresponding solution.

Contents

FAQs

What is the capacity of a serverless cluster?

Generally, a serverless cluster can handle 1 million 768-dimensional vectors. However, the actual capacity depends on your schema.

If your data exceeds the maximum capacity of a serverless cluster, please upgrade to Standard or Enterprise plan and migrate your data to a dedicated cluster. For more information about the capacity of a dedicated cluster, please refer to Select the Right CU

What can I do if I receive the error "quota exceeded[reason=disk quota exceeded, please allocate more resources"?

When inserting or upserting data, you will receive this error because your data exceeds the cluster CU capacity. To address this issue, you can follow the instructions below.

Can I change the CU type after my dedicated cluster is created?

Yes. To change the CU type, you need to follow the steps below.

  1. Create a new cluster with the desired CU type. Use the calculator to determine the CU size of this new cluster.

  2. Migrate the data from the current cluster to the new cluster you just created. Alternatively, you can also contact us to handle the data migration between clusters for you. Please specify your source cluster and target cluster when reaching out to us.

Can I change the cloud region of my cluster after it is created?

Yes. To change the cloud region of your cluster, you need to follow the steps below.

  1. Create a new cluster with the desired cloud region. Use the calculator to determine the CU size of this new cluster.

  2. Migrate the data from the current cluster to the new cluster you just created. Alternatively, you can also contact us to handle the data migration between clusters for you. Please specify your source cluster and target cluster when reaching out to us.

How can I scale down my cluster CU size?

If you need to scale down your cluster CU size, please create a new cluster with the desired CU size first. Then, migrate the data in your exsisting cluster to the new cluster.

If you require assistance when scaling down the CU size, please submit a request.

Can I deploy a serverless cluster on AWS?

No. Currently, Zilliz Cloud only supports deploying a serverless cluster on GCP. If you need to deploy a cluster on AWS, please choose the Standard or Enterprise plan.

Do serverless clusters in the starter plan support customized schema?

No. The free serverless clusters do not support customized schema. However, dynamic schema is enabled by default, meaning you can always insert data with fields that are not pre-defined. Refer to Enable Dynamic Field for more details about dynamic schema.

How can I deal with a connection timeout error when I attempt to connect to Zilliz Cloud?

To establish a connection to a Zilliz Cloud cluster, you will need to provide several relevant parameters. For instance, the PyMilvus SDK's connect method can be used as shown below:

from pymilvus import Connections

conn = Connections.connect(
alias=ALIAS,
host=HOST,
port=PORT,
user=USER,
password=PASSWORD,
timeout=30,
secure=True
)

A connection timeout error may occur in the following scenarios:

  • Poor network conditions

    To address poor network conditions, it is recommended to increase the timeout duration for the connect operation. In the above code, timeout is set to 30 seconds, meaning that the connect operation will time out if no response is received within 30 seconds after the request has been sent.

  • Incorrect connection parameters

    Zilliz Cloud clusters come with TLS enabled, so to connect successfully to your cluster, ensure that you include secure in the connect parameters and set it to true as shown in the above example. Failure to do so may result in a connection failure and a timeout error prompt.

  • Non-whitelisted local IP addresses

    If you are attempting to connect to your cluster, you also need to ensure that you have turned off any VPN/Proxy connections, obtained your public IP address (private IP addresses simply do not work), and added that IP address to the whitelist for the clusters you want to connect to.

Why can’t I connect to the cluster after the cluster is created?

You can identify the problem by following these steps:

  1. Check if the cluster status is RUNNING. You cannot connect to the cluster if the cluster is being created, deleted, or when its IP whitelist is being updated.

  2. Check if the IP address of your connection is included in the IP white list.

  3. Test the connectivity of the port by running telnet in01-(uuid).(region).vectordb.zillizcloud.com 19530. If the issue remains unsolved after all above steps are tried, please submit a request.

What can I do if I cannot connect to Zilliz Cloud with Node.js SDK?

If you fail to connect to Zilliz Cloud with the Node.js SDK, please try the following:

  1. Ensure you installed the latest version of Node.js SDK.

  2. Ensure you initialize the client correctly.

    const client = new MilvusClient('https://your-db-address-with-port', true, 'your-db-user', 'your-db-pasword');

What happens to my inactive clusters?

Your serverless clusters are automatically suspended with notice after 7 days of inactivity. You can always resume the clusters when necessary. However, your dedicated clusters will not be automatically suspended due to prolonged inactivity. To save costs, we advise you to manually suspend your dedicated clusters.

Will I be charged if I suspend my cluster?

When your cluster is suspended, you will only be charged for storage, not computing. For more details about storage costs, see Pricing.