Difference between revisions of "Plugin:Disk"
From collectd Wiki
(Described disk plugin metrics which appeared in 5.5) |
(Added link to Linux kernel documentation) |
||
Line 18: | Line 18: | ||
* <em>“weighted_io_time”</em> - measure of both I/O completion time and the backlog that may be accumulating. | * <em>“weighted_io_time”</em> - measure of both I/O completion time and the backlog that may be accumulating. | ||
* <em>“pending_operations”</em> - shows queue size of pending I/O operations. | * <em>“pending_operations”</em> - shows queue size of pending I/O operations. | ||
+ | |||
+ | For details about these metrics you can also read [https://www.kernel.org/doc/Documentation/iostats.txt kernel documentation] (Explanations of fields "Field 9", "Field 10" and "Field 11".). | ||
== Synopsis == | == Synopsis == |
Revision as of 15:57, 21 February 2016
Disk plugin | |
---|---|
Type: | read |
Callbacks: | config, init, read |
Status: | supported |
First version: | 1.5 |
Copyright: | 2005–2008 Florian octo Forster 2009 Manuel Sanmartin |
License: | GPLv2 |
Manpage: | collectd.conf(5) |
List of Plugins |
The Disk plugin collects performance statistics of hard-disks and, where supported, partitions. While the “octets” and “operations” are quite straight forward, the other two datasets need a little explanation:
- “merged” are the number of operations, that could be merged into other, already queued operations, i. e. one physical disk access served two or more logical operations. Of course, the higher that number, the better.
- “time” is the average time an I/O-operation took to complete. Since this is a little messy to calculate take the actual values with a grain of salt.
Since 5.5 there are also additional metrics on the Linux platform:
- “io_time” - time spent doing I/Os (ms). You can use this metric as a device load percentage.
- “weighted_io_time” - measure of both I/O completion time and the backlog that may be accumulating.
- “pending_operations” - shows queue size of pending I/O operations.
For details about these metrics you can also read kernel documentation (Explanations of fields "Field 9", "Field 10" and "Field 11".).
Synopsis
<Plugin "disk"> Disk "sda" Disk "/^hd/" IgnoreSelected false </Plugin>
Example graphs
Dependencies
- Linux
- Either
/proc/diskstats
(Linux 2.6) or/proc/partitions
(Linux 2.4)
- Either
- Darwin / Mac OS X
- Solaris
-
libkstat(3KSTAT)
-
- AIX
- libperfstat(3)
- FreeBSD
- ports/devel/libstatgrab