ESA Disk I/O Utilization Statistics Overview

Introduction

This document describes the meaning of the Disk I/O Utilization statistics for AsyncOS on the Cisco Email Security Appliance (ESA).

Disk I/O Utilization

This section provides an overview of the Disk I/O Utilization statistics on the ESA.

View the Statistics

The Disk I/O Utilization statistics can be accessed on the ESA from both the GUI and the CLI: 

• In order to view the Disk I/O Utilization statistics from the ESA GUI, navigate to Monitor > System Overview.
  
  
• In order to view the statistics via the CLI, enter the status detail command.

Disk I/O Utilization Gauge

The Disk I/O Utilization gauge is a percentage of the highest usage that the system has seen since the last reboot. This gauge does not display a reading against a scale of a known value. Rather, it displays the I/O utilization that the system has seen thus far and scales against the maximum value since the last reboot. So, if the gauge displays 100%, the system experiences the highest level of I/O utilization that is seen since boot (which might not necessarily represent 100% of the physical Disk I/O of the entire system).

In a system that has only recently been booted and sees high loads, the Disk I/O Utilization gauge can reach more than 100% as the scaling algorithm adjusts up the maximum. Over the long term, the Disk I/O Utilization tends to smooth out, but also tends to be much lower than 100%, because system peaks are unusual and tend to drive down the average Disk I/O Utilization statistics.

Disk I/O Utilization Statistics in Use

Unless your system enters into resource conservation mode, then you can safely ignore these statistics. If you see that Disk I/O Utilization runs at or near 100% for long periods of time, it indicates one of these possibilities:

• You are driving your I/O subsystem to the limit.
  
  
• You have an incredibly consistent load that does not vary much.
  
  

In order to determine which of the two possibilities you experience, monitor you system in order to verify whether it enters into resource conservation mode, which indicates that it cannot process and deliver messages as quickly as it receives them.

Note: There are other reasons besides a bandwidth-limited I/O subsystem that can cause you to enter resource conservation mode.