InfluxDB clustering is a high availability and scalability solution for InfluxDB Enterprise, allowing for two separate software processes: data nodes and meta nodes. The meta nodes expose an HTTP API that allows system operators to perform operations on the cluster like adding and removing servers, moving shards around, and other administrative tasks. Data nodes communicate with each other through a TCP Protobuf protocol, while meta nodes use the HashiCorp implementation of Raft as the underlying consensus protocol. For optimal clustering, 3 meta nodes are recommended, with an even number of data nodes that is evenly divisible by the replication factor. Meta nodes hold metadata such as node roles, databases, and retention policies, while data nodes hold raw time series data and metadata.