How do I set up AWS PrivateLink in SingleStoreDB Cloud?
You can connect from a SingleStoreDB Cloud cluster to private services and networks. Setting up an AWS PrivateLink involves actions from your end and the SingleStore Infrastructure end. To make this process seamless, before reaching out to SingleStore Support, we recommend you get the below details and share the details with SingleStore Support.
For Inbound Request:
Inbound means that it’s the customer's apps that are querying the SingleStore,
1. Following pieces of information are required to be shared when opening a support ticket for Inbound PrivateLink:
- Cluster URL (Cluster ID).
- Mention that the request is for Inbound.
- AWS Account ID is required from the user end.
- Region Details.
2. Once the above information is shared with support, the Support team will work with the cloud Infrastructure team to get the Service Name created for your cluster, and the details will be shared with you.
3. Create an Endpoint Connection and map to the Service Name that SingleStore supplies.
4. SingleStore will then accept the request into the service to complete the connection.
For Outbound Request:
Outbound means that the SingleStore DB makes the requests (usually via pipelines, but also can be SELECT ... INTO ...), so the configuration starts from the customer’s end.
Following pieces of information are required to be shared when opening a support ticket for Outbound PrivateLink, click here to learn about steps needed to be followed on your end (Set up an endpoint service for the service you plan to expose to SingleStore) and then share the below details with the Support:
- Cluster URL (Cluster ID) or Workspace ID (For workspace clusters).
- Mention that the request is for Outbound.
- Request AWS account ID from SingleStore.
- Region Details.
- Specify if outbound to Kafka if it is your own (self-managed) Kafka cluster or a 3-rd party managed Kafka (usually Amazon’s MSK or Confluent Cloud). Depending on the type of Kafka instances, you will need to do supply an extra config to the CREATE PIPELINE clause.
Note: Your workspace and endpoint service must be in the same region.
Further steps are mentioned in the docs, click here to learn more about it.
Note: For example, access from user end AWS/GCP/Azure to SingleStoreDB Cloud cluster is an Inbound Request, and Access from SingleStoreDB Cloud to customer end Kafka is an Outbound request.