Check manual page of df

Used Space and Inodes in Filesystems

Distribution Official part of Checkmk
License GPLv2
Supported Agents Linux Windows AIX Solaris OpenVMS Hpux Freebsd Netbsd Openbsd Macosx
This check measures the usage of filesystems. The usage is checked against a warning and a critical level, which can be specified in numerous ways.

Beware: on Linux and UNIX systems the filesystem might reserve a certain amount for root (typical is 5%). This checks considers that reserved space as used. This is consistent with the percentage-column in the output of df on most distributions. So your filesystem might be at 100% in a situation where root still has 5% free space available. On some distributions, df seems to use the user allocatable space instead of the total filesystem size as base for the percentage calculation, this might result in differences between the percentage values shown by that df version and the value shown in Checkmk. From our point of view, the calculation of Checkmk is more accurate.

Inodes: if the agent provides a df subsection for inodes, this check measures the inodes usage. Thresholds default to (10%, 5%) remaining inodes and can be set/changed in ruleset filesystems.

Trends: This checks supports filesystem trends. This means that the df check is able to compute the change of the used space over the time and can make a forecast into the future. It can estimate the time when the filesystem will be full.

In the default configuration the check will compute the trend based on the data of the last 24 hours using an exponential moving average that gives more recent data a higher weight. Also data beyond the 24 hours will to some small degree be reflected in the computation. The advantage of this algorithm is an implementation, which does not need any access to any RRDs or similar storage.

Please note that when a filesystem is started to be monitored, the trend of the past is unknown. It will take at least one trend range of time until the trend stabilises and approximately reflects the reality.

Grouping: In some situations you do not want to monitor a single filesystem but a group of filesystems forming a pool. Only the total usage of the pool is of interest. The df check supports pools by defining groups. For each group you specify a name and a list of globbing patterns (path patterns containing *, ? and [...]). The name is being used as the check item. All filesystems that match one of the patterns are part of the pool. You can specify both patterns for including and excluding filesystems from a group.

When using auto discovery you specify the groups with the ruleset "Filesystem Discovery". When configuring manual checks, you specify the list of patterns in the check parameters "patterns_include" (for including filesystems in the group) and "patterns_exclude" (for excluding filesystems from the group).

Item

The mount point of the filesystem (UNIX) or the drive letter in upper case followed by a colon (Windows). For groups the item is the name of the group.

Discovery

One service is created for each filesystem the agent reports except filesystem types listed in {inventory_df_exclude_fs}. The Windows agent only reports fixed disks. The Linux agent reports filesystems that have a size and are not of type smbfs, tmpfs, cifs or nfs. If filesystem groups are configured in the ruleset "Filesystem Discovery" and a found filesystem is matching one of the patterns of a group, then instead of a service of the single filesystem one service for the group is created. The service is the name of that group in that case.