Global load balancing distributes application traffic across service endpoints in multiple regions or environments. It can route all traffic to a passive backup during a failure or divide requests between active regions. section 9.1.2
Use weighted backends to manage reliability during infrastructure changes:
- Create a separately named passive environment in another region with infrastructure as code. section 9.1.2
- Set the active backend weight to zero and the passive backend weight to 100 to fail over production traffic. section 9.1.2
- Debug the failed active environment while the passive environment continues serving requests. section 9.1.2
- Refactor toward active-active operation when both regions should process requests and tolerate a regional failure. section 9.1.2
