CFObjective: Day 2 - Scorpio Server Monitoring

Adam Wayne Lehman (not Jason Delmore)

Overview of Server Monitoring and Performance.

Wow!  Just tons of great features.  Not much to say here other than the tool gives you everything you could want.  This kind of tool has been missing from CF from day one.

Can see actual, individual sessions
Can abort processes manually (if they are growing and not ending)
Can see all variables being generated
Can see complete memory usage and what it consists of
Can call Garbage Collection manually at any time!

Requests by Memory Usage gives you a great idea of capacity.  Can see how much memory (on average) a request requires.  Simple math will tell you about how many requests your server can handle at a given moment.

Database monitoring will be incredibly useful.  Will show queries as they run, how much memory they need, and their average run time.

Neat!  Can define many different kinds of Alerts that will fire if certain things happen on your server.  The Alert won't just email a person, but also automatically (if you desire):

• Create a snapshot of the current server environment and dump to a text file
• Kill all threads running more than 60 seconds
• Reject any new requests
• Run a custom Alert CFC will all kinds of custom logic!  So you can really make Alerts do anything you desire.
• JVM Alert: Can automatically perform garbage collection at a certain threshold.  (This should really help us at work!)

There is a bit of a drawback.  Just like debugging, when server monitoring is up and running, there is about a 15% overhead on your server.  However, there is an alternate view called Multi-Server monitoring that looks at all running servers in a very low level fashion.  Then, if needed, you can drill in.