Having several databases on one machine and all datafiles from all databases are located in the same ASM, but in different disk groups. One diskgroup has now a utilization of 86.6 %, which triggers an alarm from all databases, which is incorrect as only one database has a high disk utilization.
Querying the view v$asm_diskgroup without the where condition "state='CONNECTED'". Each database sees all ASM diskgroups, but worked only with the connected diskgroup.
The alarm is triggered only at the ASM agent level (for the ASM instance) and not for the correlated Oracle instance(s). This is the right way this type of alarms should be triggered - at the ASM instance level and not at the Oracle instance level