In Splunk IT Service Intelligence, why does KPI using Adhocsearch give different result than KPI using generated search

[AnushaJone]

Hello,

I am using the Adhoc search below

index=indexname tag=oshost host=hostname

| timechart span=15min avg(cpu_load_percent) BY host

above adhoc search is giving me what I want in the result but 

 the generated search below is not providing me the expected result.

index=indexname tag=oshost host=hostname
| timechart span=15min avg(cpu_load_percent) BY host | `aggregate_raw_into_service(avg, cpu_load_percent)` | `assess_severity(c86ca62b-1055-4aee-8604-bd52d190a4c5, ea57c8b71c4ced13b8c6eedf, true, true, true)` | eval kpi="Control-M service monitoring KPI 7", urgency="5", alert_period="5", serviceid="c86ca62b-1055-4aee-8604-bd52d190a4c5" | `assess_urgency` | `gettime`

I also tried creating the base search with the adhoc query provided above, and configured Threshold Field name as cpu_load_percent. But it's giving N/A on service analyzer and not even getting the results in Aggregate Threshold Values. I am happy to provide more details!

Thanks

[skramp]

You just have to take the first part of your search, index=indexname tag=oshost host=hostname . The aggregation is done while you are doing your settings in the wizard. That means, your interval is 15mins, you split by host, you calculate an average for the cpu and you also have to define how what you want to see as aggregated information for the KPI itself.