For the latest, most advanced host monitoring, check out New Relic Infrastructure.
The Disk I/O % utilization measures the amount of time that a device is busy vs. idle. It is not based on throughput or IOPS. To view information about disk performance and IO utilization: From the New Relic menu bar, select Servers > (selected server) > Disks.
Disks page details
The Disks page lists your server's disks, along with a breakdown of free space and I/O load for each disk. Select a disk from the list to view a detailed breakdown of I/O utilization, I/O rate (MB/s), I/O operations per second, and estimated Space usage over time.
I/O utilization measures the amount of time your disk is busy and idle (for example, 70% utilization indicates that your disk is busy 70% of the time). Space usage indicates how much space might be in use seven days from now, and provides an estimate of when the disk might be full. The 7 day trend reports the average daily space change over the last 7 days.
Linux and extra disk space
Servers for Linux reports more disk space than you will see if you run the df command as an unprivileged user. This is because the operating system reserves 5% of the available blocks on each partition for root usage (in most Linux installations). Servers for Linux is reporting the true amount of available space, including the reserved blocks.
To verify how many blocks are reserved on your file system, run df as root. (Servers for Linux does not need to be run as root to collect accurate metrics).
Caution: Unprivileged users and processes can exhaust their available space before the Servers for Linux will report 100% usage, so you should consider setting your alert thresholds below 95%.
Support for NFS disks
Servers for Linux includes client-side support for NFS v2/v3 mounted volumes. NFS v4 mounts are not currently supported.
Client-side support is useful, because NFS servers are typically black boxes on which you cannot install monitoring software (such as Servers for Linux). With client-side NFS statistics, a system administrator can quickly see which filesystems are in use and important to the host that is running Servers for Linux.
However, Servers for Linux cannot retrieve underlying hardware metrics for the physical volumes that contain an NFS volume. Some of the metrics reported are synthesized from the limited information that is available to an NFS client.
The most important synthesized metrics are for I/O utilization. These are calculated by comparing the number of bytes read from/written to the NFS volume to the previous sample.
- The time and queue time metrics are queried from the NFS client. This means that the I/O utilization graphs are most useful for performance trending and not for their raw values.
- Raw metric values can be useful in determining exactly how much traffic is being dedicated to NFS. The trend represented by the charts is likely to be more valuable. For example, a sudden spike in the graph could indicate a runaway process, or a sudden drop could represent a crashed application.
For more help
Additional documentation resources include Using the Servers UI (overview of standard functions available with New Relic Servers user interface).