The Distribution Latency (Seconds) alert tracks the amount of time, in seconds, that a replication transaction is at the Publisher waiting for a Distributor to receive it, and has exceeded the warning threshold.
Causes of high distribution latency
The most common cause of high distribution latency is when a transaction has a large amount of commands. Once the transaction is complete, distribution latency decreases.
The Replication view on the Services tab displays a graph with the non-distributed count and the non-distributed time. Graphs that trend upward may indicate a problem that needs resolution.
Reduce distribution latency
You can reduce distribution latency by:
- Reducing traffic on the publisher
- Limiting non-replication related work by the publisher
- Increasing the network bandwidth available to the publisher