If you’re a SQL Server DBA, keeping your eye on the right performance metrics can mean the difference between a smooth-running system and one that’s constantly under pressure. Here are the top 10 SQL Server performance counters every DBA should monitor — with a brief explanation for each.
1. SQLServer:Buffer Manager – Page life expectancy (PLE)
What it does: Measures how long a data page stays in memory.
Healthy value: Ideally above 300 seconds.
Low values may indicate memory pressure.
2. SQLServer:SQL Statistics – Batch Requests/sec
Why it matters: Shows how many SQL queries are processed per second.
A good indicator of overall server activity and load.
3. SQLServer:Buffer Manager – Buffer cache hit ratio
Measures how often requested data is served from memory (not disk).
High ratio (95%+) is good; low means disk I/O bottlenecks.
4. SQLServer:General Statistics – User Connections
Tracks active connections to the SQL Server.
Sudden spikes may signal application issues or load testing.
5. SQLServer:Access Methods – Full Scans/sec
Tells how often full table/index scans occur.
High values may indicate missing indexes or bad queries.
6. SQLServer:Locks – Lock Waits/sec
Measures how often SQL Server processes have to wait for locks.
High values indicate blocking or deadlocks.
7. SQLServer:SQL Statistics – Compilations/sec
Monitors how often SQL Server compiles execution plans.
Frequent recompilations hurt performance.
8. Process – % Processor Time (sqlservr.exe)
Checks how much CPU SQL Server is using.
Sustained high CPU can indicate inefficient queries or lack of indexes.
9. SQLServer:Memory Manager – Memory Grants Pending
Shows if queries are waiting for memory.
Anything more than zero is a red flag.
10. SQLServer:Wait Statistics (DMVs)
Not a counter, but a DMV-based view that helps identify bottlenecks.
Query example:
SELECT wait_type, wait_time_ms FROM sys.dm_os_wait_stats ORDER BY wait_time_ms DESC;
As a DBA, these metrics help you:
*Proactively identify issues
*Optimize server performance
*Avoid costly downtime
You can monitor them using Performance Monitor, DMVs, or tools like Redgate, SolarWinds, or even Power BI dashboards(which are already explained on DataSkills).