- Defined in:
An abstract routing policy.
See subclasses for concrete implementations:
- - Read/write requests may be routed to any cluster in the instance and will fail over to another cluster in the event of transient errors or delays. Choosing this option sacrifices read-your-writes consistency to improve availability.
cluster_idand optional field
routes all read/write requests to a specific cluster. This option
preserves read-your-writes consistency but does not improve
availability. Value contains
Direct Known Subclasses