Collect a HTTP traffic capture with Fiddler or your web browser

You can collect an HTTP traffic capture two ways: either with a Fiddler capture or a web browser capture. HTTP traffic from a Java client can be captured in a Fiddler capture, but not from a web browser capture. Due to this, depending on the situation and problem, one may be more preferable than the other. IBM Support will let you know which one to capture.

 

Collecting a Fiddler HTTP trace (Windows only)

Fiddler sets up a proxy that collects all HTTP traffic on a system. This allows it to capture any HTTP calls, including Java client calls that are not directly displayed to the browser. However, if there is a proxy already setup in your environment, Fiddler can interfere with its behavior leading to unexpected results. Let IBM Support know if you have a proxy setup before attempting the Fiddler capture and if the behavior changes while collecting the Fiddler capture.

To perform a Fiddler trace:

1. Download and install Fiddler from the Fiddler website.
2. Open Fiddler and enable HTTPS decryption by going to Tools->Fiddler Options->HTTPS, and enabling the "Decrypt HTTPS traffic" checkbox. Please note that you may encounter certificate security errors when this is set, that is expected behavior.
   Note: If you need capture the HTTPS connection setup, disable the "Decrypt HTTPS traffic" checkbox and make sure the "Capture HTTP CONNECTs" checkbox is enabled from Fiddler Options.
3. With Fiddler open, make sure the capture is enabled by checking if File->Capture Traffic is enabled. Then re-create your issue while Fiddler is open.
4. Save your Fiddler capture by going to File->Save->All Sessions 
   Save the File as a .saz and provide that file to IBM Support. Let IBM Support know of any time differences between server and client along a time stamp of when you re-created the issue.

For an example of collecting a Fiddler capture, we have a video walkthrough showing how to collect a Fiddler capture on Youtube.

 

Collecting a browser traffic capture

A web browser captures will collect all of the http traffic directed to the browser while it is enabled. While it may miss any non-browser HTTP calls (such as Java client calls), you generally do not need to install any new applications to collect it.
Note: Web browser captures will also include browser cached responses, so for web browser specific issues a browser capture may be more preferable than a Fiddler capture.

Collecting a browser capture depends on which web browser you are using. We have a video that details the different ways of capturing browser traffic for Internet Explorer, Firefox, and Chrome.


After you are done collecting the HTTP capture, upload the data to the support case or upload via ECURep. Please see this link for details.