You can filter the metric builtin:host.cpu.system with host you are looking for and then pin it to the dashboard
Not sure if it helps, but your screenshot has Processes CPU Consumption, not Services.
Processes CPU consumption (and other metrics per Process) can be found under Technologies > Generic > CPU (or just put builtin:tech.generic.cpu.usage im Metric field)
@MAkimov There isn't a way to target the CPU consumption of a Service as a dashboard Tile/Chart. You can tile/Chart the next level down which would be done by marking things as Key Requests. Keep in mind, marking key requests are always forward facing.
I would recommend putting in a RFE as the functionality that you are looking for does not exist in the product as of today from a dashboard view point.
As a work around, you can use a mark down tile and drop a link to the page that you want to access. But even then it wont show you values on the dashboard, rather just quick and easy access to the page.
Maybe you can try this metric expression. Maybe it is a completly worng way of thinking. Try it.
(builtin:tech.generic.cpu.usage:filter(and(or(in("dt.entity.process_group_instance",entitySelector("type(process_group_instance),entityName.equals(~"YOUR PG INSTANCE NAME~")"))))):splitBy():avg)
*(100*(builtin:service.cpu.time:filter(and(or(in("dt.entity.service",entitySelector("type(service),entityName.contains(~"YOUR SERVICE NAME~")"))))):splitBy():avg)
/builtin:tech.generic.cpu.groupTotalTime:filter(and(or(in("dt.entity.process_group",entitySelector("type(process_group),entityName.equals(~"YOUR PG NAME~")"))))):splitBy():avg)
Last column is the above mentioned metric expression:
3rd column came from the first two with a % calculation, then I counted the 10,2% of the 21,59% which is the fifth column.
The above calculated process CPU detailed breakdown with the test service below. I know that 2,68 % is not 2,2% but it is quite close. I have tested it with randomly and the results are really quite close the CPU detailed breakdown values:
I hope it helps.
Also thought about utilizing builtin:tech.generic.cpu.usage metric but @MAkimov asked to see the CPU% per service while this metric is showing the process CPU% and one process can run more then one service hence IMO its good answer but not for @MAkimov request 🤔.