In steady state replication, this duration indicates replication lag, but the secondary will still be able to serve reads. In the startup recovery case, we will not continue with startup until the index build is completed, so this will indicate loss of availability. It would be beneficial to differentiate between these two cases.
- related to
-
SERVER-132933 [v7.0] Wait for secondary to reach SECONDARY after restart in index_build_blocking_commit_duration_metrics.js
-
- In Code Review
-