cancel
Showing results for 
Search instead for 
Did you mean: 
Create Post
Level 8

metric results accuracy

Hello,

I am using ignite DPA for past 3 days for monitoring a sqlserver 2008 R2 which hasnt been rebooted in an year. Application seems to be running slow for the past few days and I am currently looking at the perf metrics like

sql disk read latency, sql disk write latency, sql compilations per sec and sql recomiplations per sec.

Per ignite the avg sql disk read latency is shown as 90 ms,

avg sql disk write latency as 140ms ,

avg sql compilations per sec as 300

avg sql recomiplations per sec as 35.

Can these be taken into consideration for improving performance or do you think the server needs to be rebooted before monitoring using DPA?

0 Kudos
4 Replies

There should be no relationship between reboots and the collected metrics.  If you say the application is running slowly and are seeing disk latency in the 90-140 ms range then you basically already have identified an issue.  I normally expect my systems to have latencies under 20 ms, but obviously each company sets their own standards for what "fast enough" means to them.

- Marc Netterfield, Github

0 Kudos
Level 8

Agree with Paul's article. I wanted to know if the metrics shown are aggregates.

0 Kudos

They are the averages during the interval, I want to say its a 60 second avg normally for dpa

- Marc Netterfield, Github
0 Kudos