eG Help

If transaction log records are not sent quickly by the primary database or are not applied quickly by the secondary database, then the data in the primary and secondary databases will be out of sync; this will cause significant data loss during a failover. To avoid this, administrators must keep track of the log record traffic between the primary and secondary databases, proactively detect potential slowness in synchronization, figure out the probable source of the bottleneck, and clear it to ensure proper synchronization between the primary and secondary databases. This is where the MsSqlAlsNetLncyTest test helps.

This test measures the rate at which transaction log data is sent to the secondary database for synchronization on each SQL server instance, and the time taken by the secondary database to apply the data. In the process, the test pinpoints bottlenecks in database synchronization and where exactly the bottlenecks lie.

Measurement	Description	Measurement Unit	Interpretation
Send_to_replica	Indicates the amount of data (in bytes) sent from the primary availability replica to the secondary availability replica per second during the last measurement period.	KB/sec
Send_to_transport	ndicates the amount of data (in bytes) sent over the network from the primary availability replica to the secondary availability replica per second during the last measurement period.	KB/sec
Flow_control_time	Indicates the time duration for which the log stream messages were waiting in the Flow Control mode per second.	Msecs/sec	Ideally, the value of this measure should be low. A gradual/sudden increase in this measure indicates that the network over which the log messages are sent is experiencing slowdowns/network delays and noise. A high value for this measure is also indicative of potential data loss which is much more than the estimated Recovery Point Objective (RPO).
flow_control_per_sec	Indicates the number of times Flow Control mode was initiated per second.	Waits/sec	A high value for this measure indicates that the network is congested and is experiencing slowdowns.
Avg_flowcontrol_time	Indicates the average time the log messages should wait in the Flow Control mode.	Secs/wait	This measure is a ratio of the Flow_control_time and the flow_control_per_sec measures. A low value is desired for this measure.
Resent_messages	Indicates the number of Always on messages i.e., log stream messages that were resent over the network during the last measurement period.	Number	Ideally, the value of this measure should be low. A high value for this measure is a cause of concern as this indicates a high network latency or network congestion or network noise.