Apparently, the Splunk OS TAs don't capture time and if there are index time delays, _time
would be skewed and actually be _indextime
.
For example, the output of df.sh
is -
Filesystem Type Size Used Avail UsePct MountedOn
/ xxx 50G 18G 30G 37% /
/yyyyy xxx 600G 401G 186G 69% /yyyyy
/zzzzz xxx 50G 18G 30G 37% /zzzzz
Is there anything we can do about it?
Hello @danielbb,
I think TA_nix developers did it intentionally - the disk usage dosn't change rapidly and any change will not be apparent because of using the human readable format anyway - the usage is rounded to GB.
But it is easy to change by modifing df.sh (prepend with date command) and adjusting props.conf on the indexer side.
For a real time statistic it is better to use sysstat, particularly
sar -F 1 /dev/sda1
or
sar -F 1