Tetrate Service BridgeVersion: 1.10.x
Operating and Testing HA and Failover
This document explains how you might test failover, and operationalize tasks such as draining and restoring clusters
A platform operator may need to manually:
- Take a Workload Cluster out-of-rotation, allowing existing requests to complete, before performing maintenance on that cluster
- Take an Edge Gateway out-of-rotation, allowing cached DNS entries to time-out, before performing maintenance on that Edge Gateway
- Define a region as ‘active’ or ‘passive’ (the default Tetrate model is all-active)
The worked examples in the Workload Cluster failover and Edge Gateway failover guides illustrate various ways to take components out-of-service in a controlled and predictable manner, and the best implementations will be influenced by specific topologies and choice of GSLB solutions.