Description
This alert will calculate the average CPU Utilization in the last 5 minutes based on Ignite CPU metric data collected once every minute by default.
Note: The example below uses the "CPU Utilization" metric, but a very similar query can be used for other metrics as well. For example, if you wanted to monitor O/S Memory Utilization, you could modify the "where mn.NAME = 'CPU Utilization'" line to "where mn.NAME = 'Memory Utilization'". For a list of all metric names, you can run these queries:
select id, name from cond; -- use the ID for the instance in the query below for <DBID>
select name from con_metrics_names_<DBID> order by name; -- substitute the value of name into the alert query
ALERT DEFINITION
To create the alert, click on Alerts > Manage Alerts tab and create a Custom Alert of type Custom SQL Alert - Multiple Numeric Return. Configure the alert with values similar to these:
Alert Name: CPU Utilization
Execution Interval: 5 Minutes
Notification Text: The average CPU utilization for the last 5 minutes has exceeded thresholds
SQL Statement:
SELECT NVL(AVG(md.V*1.0),-1) avg_cpu_5min
FROM con_metrics_names_#DBID# mn
INNER JOIN con_metrics_#DBID# m ON m.metric_name_id = mn.id
INNER JOIN con_metrics_detail_#DBID# md ON m.id = md.metrics_id
WHERE mn.name = 'CPU Utilization'
AND md.d >= SYSDATE - (5/1440)
Execute Against: Repository
Units: % CPU Utilization
High Threshold: Min 90, Max empty
Medium Threshold: Min 80, Max 90