I'm fairly new to Splunk and have a search that basically returns a count of the number of times a device logs in to our system and uploads data each week. The time chart looks similar to this.
_time Device A Device B Device C Device D
2015-10-04 1 1 1 0
2015-10-11 1 1 1 0
2015-10-18 1 1 1 2
2015-10-25 1 0 1 1
2015-11-01 1 0 2 1
2015-11-08 1 1 1 1
2015-11-15 1 1 3 1
The only devices I'm concerned about are those that have zero connections at some point: Devices B and D. How would I filter those that are working as intended, Devices A and C, from my results?
The total device list can be in the thousands depending on the geography I search in. I'm only interested in the ones that appear to be having issues, those with a zero count for a week or more, so I can focus on that population.
Thank you in advance for any help.
This should do:
base search | timechart count by device | untable _time device count
| eventstats min(count) as min by device | where min=0
| xyseries _time device count
First, make your initial timechart and flip that into a stats-like table, then filter by "device has a value that's zero", then flip back into timechart-like table.
This should do:
base search | timechart count by device | untable _time device count
| eventstats min(count) as min by device | where min=0
| xyseries _time device count
First, make your initial timechart and flip that into a stats-like table, then filter by "device has a value that's zero", then flip back into timechart-like table.
That worked perfectly, thank you for the step by step explanation, it was very helpful.