Comment by richieartoul
2 years ago
Requests to the sequencer are batched heavily. If the sequencer fails, the cluster goes through a recovery and will be unavailable for 2-3 seconds and then recover.
2 years ago
Requests to the sequencer are batched heavily. If the sequencer fails, the cluster goes through a recovery and will be unavailable for 2-3 seconds and then recover.
Good point about the batching! Any idea what kind of ReadVersion qps throughput you can get this way? And yeah, 2-3s unavailability seems fine.