uberAgent - IOPercentDiskTime metric shows value > 100%
Article ID: CTX695838
Updated On:
Description
The difference between metrics from older and new agents is expected. In the older version, max value was 100%.
From release notes:
Performance counters [I1225]: percent values are not capped at 100 anymore.
uberAgent is collecting correct metrics now.
In case of very high values, admin should check the machine with high IOPercentDiskTime.
You can query performance counters from Powershell
Get-Counter -Counter "\PhysicalDisk(_Total)\Avg. Disk Queue Length"
Get-Counter -Counter "\PhysicalDisk(_Total)\Current Disk Queue Length"
Get-Counter -Counter "\PhysicalDisk(_Total)\% Disk Time"
You can also query using the Perfmon Tool (counter: \\PhysicalDisk(_Total)\\% Disk Time).
Issue/Introduction
Splunk dashboards and queries show that the IOPercentDiskTime metric is above 100%.
Query:
Index=uberAgent sourcetype=uberAgent:System:SystemPerformanceSummary2 host="affected machine" | timechart max(IOPercentDiskTime)
Not seeing the same behavior when checking metrics from the machines running older uberAgent versions.
Additional Information
High Disk time% can appear when:
- there is a storage issue
- there is a high number of IO
- 100% RAM utilization on the machine and pagfile havily used
Was this article helpful?
Yes
No
Powered by
