topic Re: Working with loadjob in Reporting

Working with loadjob

daktapaal — Sat, 01 Feb 2014 02:37:13 GMT

Hi, I am hoping you can help me here.
I am running a search out of a saved search using the load job.. I did something like :
|loadjob savedsearch="abc:search:my search"

This produces a table of field cols that look like host1#maxpings , host2#maxpings etc.. Where max pings s the maximum pings that the host can have. Under these columns are the daily pingNumbers for these hosts.
something like :
…...

-………… host1#23…………. host2#56.1

day1 ………...3………………………..4
day2 ………..10……………………..11
day3 ………..20……………………….50

I need to find out if the average pings for 3 days is more than the numbering the column name. for example : in the above, I need to find for column1, (3+10+20)/3 < 23 and for column2 (4+11+50)/3 < 56.1 … I should then show only those columns, where the avg number is less than the number in the column…
i want to pseudocode something like |loadjob savedsearch="abc:search:my search" | stats avg(*) as average(*) | where average < substring-after( col-name,'#') . So that this will show only those columns where the average is less than the number in the column.
I am losing hopes on the help from google and splunk docs. Help with this will be highly appreciated.
Thanks in advance.
dT

Re: Working with loadjob

martin_mueller — Sat, 01 Feb 2014 10:26:22 GMT

I've recreated your table like this:

| gentimes start=-1 increment=2h | eval day=1 | accum day | eval day="day".day | eval host0#10=random()%50 | eval host1#20=random()%50 | eval host2#30=random()%50 | eval host3#40=random()%50 | eval host4#50=random()%50 | fields day host*

Based on that, this may be the postprocessing you're looking for:

... | stats avg(*) as * | transpose | rename "row 1" as average column as host | eval limit=replace(host, "^.*?#", "") | where average < limit

Compute averages, transpose and rename results, select part after "#", compare with average.

Re: Working with loadjob

Rmddas — Sat, 01 Feb 2014 19:07:53 GMT

cool. thanks.I figured this out last night 🙂