Has this issue ever been addressed? I am having similar situations in 12c Release 3. Metrics get populated in the charts of generic and aggregate services...MOST of the time, but a lot of the times they are inaccurate. I have a collection of 3 weblogic servers put together as a generic system/service. All 3 are up and running, yet the generic service Usage metric I created to look at the "Status" of all 3 servers in a SUM, is still showing 2. It shows the 3 servers as up with no alerts, but the metric only reads 2 on the chart and takes around an hour to update. The Generic system and the Weblogic targets themselves show an UP status just fine...but this specific metric is wrong. I have tried to recreate it and it did not change it.