I am using Canary Response Time, Query Monitor, Productivity, System Health with couple of other portlets. These are good in what these are ment to do. What I want is to get info about any thing new, if it happens on the system.
if a user used more sessions, IO, CPU, etc than it uses normally. OR there is some new activity on the system.
In short I just want to get a portlet which can give me quick info about the cause if system behaves degraded. (I think any portlet accessing DBQL history data will surve this purpose)
As Dieter suggests, setting up alerts and then using the Alert Viewer portlet would be a good way to track irregular system behavior. You could even setup email notifications as an associated alert action.
Other portlets to consider would be the trending portlets, Metrics Graph and Capacity Heatmap, where you can monitor overall CPU usage or other metrics to see where system usage increases.
Query Spotlight allows management of queries of concern over durations of time so this should definately be part of your analysis tools.
Also consider the Todays Statistics portlet for easy comparison of what happens today against the same time frame a week ago.
Regarding the Metrics Analysis Graph, can you clarify for me what the Y-axis of the yellow AWT graph line tells us? The value for instance has bounced between 14 and 0.13 over the past couple hours. When I hover over the line at any point, it shows the value and time but also shows a Min and Max values. Can you explain what these values are telling me and what the min and max values mean? Is this the number of AWTs in use on a single AMP or the average of AMPs? Is this the number of AWTs in use or available? Is it for a given AMP or the system as a whole. Thanks so much.
In Metrics Analysis, the AWT metric shows the average number of AWTs in use per AMP on the system. This is the solid line in the graph. The min value in the balloon represents the least number of AWTs in use across all AMPs on the system during that particular interval. The max value represents the highest number of AWTs in use across all AMPs on the system during that particular interval. You will also see the range between the min and max values with a light shading when you select just the AWT metric on the graph.
Also keep in mind that Metrics Analysis can serve as both a monitoring and trend analysis portlet. When you are looking at the 1 hour range, you are likely seeing each point on the graph represent a single collection point and therefore a snapshot of your system at that point in time. As you move into the larger ranges, the portlet becomes more of a trend analysis portlet and the data at each point represents minutes or hours worth of data. As you start to move into these longer ranges, the average is still an average over a specific time interval. However, the min and max still represent the true observed min and max during that interval (i.e. these are not the min and max average).
Hope that helps!