What are System Metrics?
System Metrics are kind of an extended status report for your server. FusionReactor stores information over time such as how many pages returned the status code 304, how many pages were aborted by crash protection and which pages have taken the longest to execute.
How do I get an overview of the System Metrics?
From the FusionReactor Administrator, click "System Metrics" from the table of contents (Metrics section) and you will see the Request Metrics page. It will look something like this:
At the top of the screen shot you will see the refresh bar. You can use this to manually or automatically update the running requests listed below. Below you can see six metrics graphs which are a new feature introduced in FusionReactor 3.0. The graphs are a Flex component that show different and detailed visual metric information for:
- Request Activity
- Average Request Time
- JDBC Request Activity
- Average JDBC Time
- Memory Usage
- CPU Usage
The six graphs are interchangeable allowing a distinct metric graph to be selected to show a larger and more clear view, for example you can click on any graph to maximize it and click on a maximized graph to return to the standard view. You can also flip one or all of the graphs between the minute and hour view by selecting one of the options from the little clock menu at the top right corner of each graph.
Below the graphs, the page itself lists all of the HTTP return codes which have been produced by completed requests, the number of pages aborted by crash protection, the amount of time the server has been up, the number of requests serviced, average request time and current memory settings. If you click on a HTTP return code or crash protection type then you will be shown the requests which fall into that category (as long as they are still in the request history.)
How do I see the longest running requests?
From the FusionReactor Administrator, click "Longest Requests" from the table of contents (Metrics section) and you will see the Longest Requests screen. It will look something like this:
At the top of the screen shot you will see the refresh bar. You can use this to manually or automatically update the running requests listed below. Over time you will eventually only see very long running requests on this page so you can reset it at any time by clicking the "Reset Longest" button which is located above the request list.
How do I see the slow running requests?
From the FusionReactor Administrator, click "Slow Requests" from the table of contents (Metrics section) and you will see the Slow Requests screen. It will look something like this:
At the top of the screen shot you will see the refresh bar. You can use this to manually or automatically update the running requests listed below. This will show any requests which have been running long enough to be considered slow. A slow request is simply one which has been executing more than the threshold value specified in the metrics settings.
How do I see the longest JDBC requests?
From the FusionReactor Administrator, click "Longest JDBC" from the table of contents (Metrics section) and you will see the Longest JDBC Requests screen. It will look something like this:
At the top of the screen shot you will see the refresh bar: 
. You can use this to manually or automatically update the running JDBC requests listed below. Over time you will eventually only see very long running JDBC requests on this page so you can reset it at any time by clicking the "Reset Longest" button which is located above the request list.
How do I configure metrics?
From the FusionReactor Administrator, click "Metrics Settings" from the table of contents (Metrics section) and you will see the Metrics Settings screen. It will look something like this:
From here you can change some of the Metrics behavior:
Enable Metrics: Enable or Disable the whole metrics system.
History Size (Slow and Longest Running Requests): This is the number of requests which will be shown on the Longest Running and Slow Requests pages.
Slow Request Threshold: Enter a number of seconds here and anything which runs for longer than this specified time will be considered a slow request.
History Size (Longest Running JDBC Requests): This is the number of requests which will be shown on the Longest JDBC Requests page.
Why would I want to see Metrics?
Metrics are a great way to get a quick idea of how your server is running. You can see if you are getting a lot of page breaks or server errors and you can see the pages which caused these problems. Even if all pages are completing, you can see if they are running fast enough. Basically, it tells you how healthy your application is and where you should look to improve it.
