+- `clusterId` is no longer derived automatically, and applications that connect to more than one database cluster must now set it explicitly. Previously the wrapper generated a cluster id from the connection URL and, for AWS RDS clusters, converged connections made through instance endpoints, IP addresses or custom domains onto the cluster endpoint's id. That derivation, along with the suggested and primary cluster id caches, has been removed: `clusterId` is now taken only from the configuration parameter and defaults to `1` for every connection. An application that connects to several clusters without setting a distinct `clusterId` per cluster will have all of them share the same topology cache and monitor entries, which leads to incorrect topology and failover to the wrong cluster. Give every cluster its own value, and use the same value for every connection to a given cluster. See [Cluster ID](https://github.com/aws/aws-advanced-nodejs-wrapper/blob/main/docs/using-the-nodejs-wrapper/ClusterId.md).
0 commit comments