The elapsed time between a database operation being initiated and its completion, a key performance indicator for SQL Server health that encompasses query execution time, I/O response time, and network round-trip time.
- DBAs monitor multiple types of latency: Query latency (time for a query to return results), I/O latency (time for storage to service a read or write request), and network latency (round-trip time between client and server).
- Storage I/O latency is one of the most impactful performance factors: Production SQL Server systems should target less than 5 milliseconds for data file reads and less than 1 millisecond for log file writes.
- High query latency is diagnosed using wait statistics, execution plans, and Query Store: These tools help identify whether the underlying cause is CPU, I/O, memory pressure, locking, or network-related issues.
- In Always On environments, replication latency determines the effective Recovery Point Objective (RPO): It also affects whether secondary replicas are suitable for offloading read workloads.
- Relevant Idera tools: SQL Diagnostic Manager captures latency metrics across queries, disks, and networks, providing historical trending and configurable threshold alerting.
- Related terms: Wait Statistics, I/O, Query Store, Execution Plan, Always On Availability Groups.
