The FusionReactor ColdFusion Plugin is able to profile the execution of ColdFusion CFML at line level. This data can be used to identify performance issues deep inside ColdFusion pages and components and can be an invaluable tool for finding problems within your applications. This feature also gathers information about the CFML code coverage which can be very valuable when testing applications.
Please note that this feature uses line number information generated by the ColdFusion Server that is known to be inaccurate in some cases and beyond our control. For this reason we cannot guarantee 100% accuracy of the information. Also note that the performance of the ColdFusion server may be HEAVILY impacted while tracking line performance profile data. You should use FusionReactor version 6.0.4 or higher.
- You must have FusionReactor 6.0.4 or above
- Go to FusionReactor->Plugins->Active Plugins
- On the Active Plugins page choose the FusionReactor ColdFusion Plugin Configuration
- On the Configuration page - Enable Track CFML Line Execution Times
- Save the configuration - Note this will have immediate effect. When you enable CFML Line Performance Tracking the currently compiled templates have to be recompiled with instrumentation in them for collecting line metric information. This results in the ColdFusion template cache being emptied until the templates have been recompiled. Recompilation can be an expensive task and could result in an application stalling. Also note the compilation will take longer than normal with CFML Line Performance Tracking because the code must be instrumented heavily with metric gathering data.
- We have written a simple tool in CFML that allows you to examine the performance of your ColdFusion application code on a line by line basis. The tool requires that you have FusionReactor installed and have enabled Track CFML Line Execution Times in the FusionReactor ColdFusion Plugin.
This tool can be found here in our GitHub repository.
- If you don't want to use the tool - then you can also get access to the data via the FusionReactor API (FRAPI)
With tracking enabled, FusionReactor captures the information about each line of CFML code executed and stores the data internally. Using a FusionReactor API you can gain access to this information and use it generate reports about the performance of your code:
The following code will output the performance information for all of the files that have been profiled:
Considerations for Production Systems
The following points must be considered when enabling and disabling this feature on production systems:
- Enabling or disabling CFML Line Performance Tracking will result in the ColdFusion Template and Component caches being emptied, requiring a complete recompile of your application code which may be very time consuming and will impact the performance of your application abruptly and immediately
- Under production load, the ColdFusion server may not be able to recompile the code fast enough to keep up with user load, resulting in requests waiting and your application could stall completely.
- CFML Line Performance tracking can be an expensive action as every single execution of every line of code by every request/thread is tracked. Under production load, when you have a lot of requests/threads running (potentially the same code) the performance of the application may be too heavily impacted for this feature to be enabled. The granularity of the data capture that this feature requires may mean that your server does not have enough cpu capacity left to meet both the production application's load and the line performance tracking requirement at the same time.
- The overall performance of the application can be heavily impacted. The performance impact of this feature ranges typically between 2-10% but can be much higher depending on the code being executed.
- Configuring fewer simultaneous requests in ColdFusion may mitigate these affects in some cases.