Clusters represent a logical grouping of resources in JumpWire. Databases, client credentials, and policies can call be scoped to a specific cluster. This makes it very easy to separate production, staging, and dev infrastructure. Multiple clusters can even be used in the same environment for further isolation.
Databases must be configured for a specific cluster. The proxy will connect directly to the database, so any internal hostnames or port mappings should be valid and reachable from the configured cluster.
If a cluster is deleted, database configurations that were set to use it will no longer be valid. They will still exist and can be edited in the web interface to use a different cluster.
Like databases, individual policies can be assigned to a cluster. However, using a cluster is optional. By default, all policies will be shared by all clusters.
It can make sense to assign a policy to a cluster when testing a change, or for configuring different handling rules across different environments.
Each cluster has access tokens specific to it. When starting the JumpWire engine, the token provided to it will determine which cluster it connects as and what data it is capable of retrieving. Tokens can be generated and revoked at any time.