DR Failover Timing Summary
| Phase | Duration |
|---|---|
| GLB detects primary failure | 30-60 seconds |
| Cloud Function triggers (if enabled) |
~3 minutes (configurable via
dr_cloud_function_pre_failover_failure_seconds)
|
| Standby LogScale pod starts | 2-5 minutes |
| Promotion Phase 1 (generic routing) | 5-10 minutes |
| Promotion Phase 2 (dedicated routing) | 2-5 minutes |
| New node pools scale up | 5-15 minutes |
| Total RTO (Recovery Time Objective) | 15-35 minutes |
Note
The RTO range depends on cluster size, node pool scale-up time, and whether the Cloud Function pre-scales standby nodes before manual promotion.