Filesystem metrics collected by APMIA / SystemEdge do not match when we check on the servers.

Products

• CA Application Performance Management (APM / Wily / Introscope)

Issue/Introduction

We are using APMIA with SystemEdge to collect and alert on filesystem metrics for Unix servers.

We noticed that the metric value displayed in APM does not match what is actually on the server.

For example, we are monitoring the Percent Used metric -  in APM the percent used shows 86%, but when we run "df -h" on the server the percent used shows as 92%.

Please explain the discrepancy.

Environment

Release : 23.1

Resolution

There is a difference in calculating the metric value that we are reporting as per below

Percent used: The percentage of used blocks, obtained by dividing Used (Blocks) by Size (Blocks)

To verify this, run df -k command to show the storage in 1KB blocks. Then divide the values to get the percentage. This is matching with the APM reported value.