Load Balancer for ClickHouse Clusters

Load Balancer for a ClickHouse cluster is a Network Load Balancer that has the following functions: 

  • Direct client requests to one of ClickHouse server nodes in the cluster 
  • Performing regular health checks against nodes to ensure connectivity 

The decision to use a load balancer depends on your requirements and constraints: 

  • Load Balancers ensure high availability and optimized resource utilization by distributing queries across multiple nodes, thereby avoiding bottlenecks and providing a single connection endpoint for simplified client configuration but come at the additional cost of an Enterprise Feature (as well as any cost for the Load Balancer itself, payable by RIYOA customers to the cloud provider.). 
  • Load balancers may not be required for small-scale deployments, and they may not be unsuitable for latency-sensitive environments, as load balancers will increase latency by adding processing time and routing overheads. 

 

Operational properties: 

  • Connections via the Load Balancer are secured over TLS, which terminates on the target nodes in the cluster. This ensures security while direct connections to each individual node are still supported. 
  • The Load Balancer listens on the following ports and protocols: 
  • 9440 – Native protocol SSL/TLS port 
  • 8443 – HTTP SSL/TLS default port 
  • Instaclustr provides a DNS record that can be used to connect to the Load Balancer which can be found on the connection info page.

  • REST API clients do not need to support or be configured with sniffing when using Load Balancer. 
  • The Load Balancer is configured to provide high availability, specifically by selecting proper components and resource tier in the cloud service provider. 
By Instaclustr Support
Need Support?
Experiencing difficulties on the website or console?
Already have an account?
Need help with your cluster?
Contact Support
Why sign up?
To experience the ease of creating and managing clusters via the Instaclustr Console
Spin up a cluster in minutes