Casquatch is designed to support failing over database queries from one datacenter to another. This generally causes increased latency and weaker consistency in exchange for improved availability.
Casquatch APIs are designed to build statements which then get passed to execute or executeASync as appropriate. If an exception is caught then it is passed to the class defined by casquatch.failover-policy.class which implements FailoverPolicy. This interface takes a statement and an exception to determine if it should be replayed on the remote site. If so, then it is attempted again with the profile defined in casquatch.failover-policy.profile. This parameter supports being nested in additional profiles allowing chained failovers as required.