The Pipeline tab allows users to edit configuration parameters such as capture, limits, and cookie IDs.
- If the fields are left blank, the device attempts to read default values from the
You can select either of the two views to display pipeline settings:
- Pipeline Settings - Define data sessioning, time grading, capture mode, and more.
- Edit Type List - Specify the data types and file extensions to include or exclude from capture.
Note: PCA capture types must be reviewed during any installation of the CX Passive Capture Application or if the monitored web application is updated.
By default, the PCA is configured to create a single instance of the PCA pipelined process for use by all instances of the PCA application. If needed, more instances of the pipelined process can be created to more efficiently distribute the work across available resources.
When multiple instances of the pipelined process are created, each instance of the PCA application submits the TCP packets. These packets are assembled by its reassd process to the pipeline queue. This queue then distributes the packets for processing to the configured pipelined instances in a round-robin fashion.
The PCA pipelined process takes the individual HTTP hits and performs processing on them, such as dropping hits, applying privacy, data compression, and more.
The effectiveness of multiple pipelines depends on the number of available CPU cores. The general rule of thumb is to allocate one available CPU core for each pipeline instance. If four idle CPU cores are available, you must configure no more than four pipeline instances if you expect a full CPU load.
128. It is not likely you must approach this limit.
To retrieve statistics on an individual pipeline instance, run the following commands.
Instance #0 stats:
Instance #N stats:
ctcstats -I<#N> instdata
<#N> is the instance identifier.
The PCA can be configured to create a session id for each hit that is based on cookies injected into the request. When enabled, you can configure the request section and field information that is containing the session identifier.
The cookie on which to session must be unique and persist for each hit in a session.
When data sessioning is enabled, the session id/cookie specified by the
Name value is hashed to create the session identifier (
TLTSID). If the
Cookie Injector is in use, it creates this value.
- The name of the cookie on which you want to sessionize. Accepted values include
- To sessionize on multiple cookies, you can insert multiple delimited values in this field.
Accepted delimiters are comma (
,) or semi-colon (
- To sessionize on multiple cookies, you can insert multiple delimited values in this field. Accepted delimiters are comma (
- You can use this field to specify the section of the request in which to look for the
Field Name. If this value is not specified, then the entire request is searched for the field, and the first occurrence is used.
- You to sessionize on a specific part of the session cookie, as specified by
Field Name. For example, if your cookie value is 32 characters long, setting the value to
0 15sessionizes on the first 16 characters. This feature is useful if the session cookie is appended with a static string:
. jsessionid=<unique keystuff here>mycookie;
If enabled, session sampling specifies a percentage of sessions to forward to the delivery peer. The remaining sessions are deleted from the capture. Session sampling allows for statistically significant volumes of data capture without burdening the system with production volumes.
The Capture Mode setting specified the types of data that the PCA captures and drops from the capture stream. The following settings are available:
- Capture Mode
- Configures the Passive Capture software to capture only HTTP(S) request and response objects for
- Experience Analytics replay clients can acquire and load these items on an as-needed
basis during replay. Note: Business capture mode is the default and recommended setting.
- Experience Analytics replay clients can acquire and load these items on an as-needed basis during replay.
- Configures Passive Capture to capture HTTP(S) requests and responses, as well as file objects
associated with each hit. These static objects, such as image files, are captured by the PCA and
passed to the Windows pipeline for further evaluation.
Note: In BusinessIT mode, the response object payload body is not saved. Only the image file and the request are captured and processed.Note: Changing the capture mode from Business to BusinessIT significantly increases the volume of traffic that is captured, processed, and stored by Experience Analytics. Before you change, review your entire Experience Analytics solution to verify that you have the system resources to manage the payload increase. For more information, please contact Experience Analytics Professional Services.
If enabled, Time Grading can assign a grade to a hit in one of the following three areas:
Web Server Page Gen
- How long it takes the web server to serve up the page.
- Measures network speed based on how much time a packet spent on the network.
- How long it takes an arbitrary packet to travel from the client to the Web server.
Based on the three criteria, Time Grading assigns the hit a numerical grade. The default grade descriptions (Excellent, Normal, HighNormal, High) can be edited, along with their corresponding maximum values.
The following table describes the hit processing settings.
||Determines whether RawRequest is on. RawRequest is an aid in debugging. The
default value is
Note: Set this value to
||Determines whether ResponseHeaders are on. ResponseHeaders are aids in
debugging. The default value is
Note: Leave this value set to
||This option determines whether to URL-decode URL fields.|
||If this option is selected, a cookies section is added to the request.|
||If enabled, this option checks the last 100 bytes of the response body for
||When this option is selected, responses are automatically inflated in the PCA
pipeline. If a response has a content-encoding header whose value is
Note: This setting must be set to
||When a hit response is inflated and the inflate process fails, the response is replaced with HTML reporting that an inflate error is occurred. Select this option to disable the replacement HTML, which preserves the original response.|
||Select this option to deflate hit responses before they are sent to a target recipient. The hit responses are compressed by using the HTTP deflate method. Enable this option to reduce the amount of network bandwidth that is used to deliver hits to the Transport Service.|
||When selected, internationalization support is enabled on the PCA. The request
urlfields are encoded to UTF-8, and an attempt is made to detect the response body encoding.
Note: There is a known issue in which special characters submitted in JSON data from a client framework are mangled.
||When enabled, this option deletes image hits in the capture stream that meet specific criteria.|
||The largest acceptable response size (in bytes). The default value is
||If the PCA is not able to detect the encoding type of a response, this value is used.|
||If the PCA is not able to detect the encoding type of a request, this value is used.|
||This statistic reports each time a hit (either Request or Response) cannot be encoded to UTF-8 which causes the complete hit to be dropped. This scenario also generates an error (ERR) in the log.|
Delete Images on PCA Side
The PCA provides DelImages functionality in the PCA pipeline. When DelImages is enabled in the PCA, the volume of data that is captured, processed, and transferred to the Windows server is reduced.
If a hit meets the following criteria, then the hit is identified as an image hit and is dropped:
Delete Images on PCA sideis enabled.
- The file extension for the response is listed on the Excluded File Extensions list.
- Either of the following criteria is met:
StatusCode = 200or
StatusCode = 304)
The previous criterion determines if the PCA is in BusinessIT mode (CaptureType=3) and an image request is returned with an "okay" status or a Not Modified status (StatusCode =304).
HTTP_USER_AGENT contains "RealiTeaViewer"or
HTTP_USER_AGENT contains "TeaLeafFileGetter"
The previous criterion determines if the request is being made by CX RealiTea Viewer desktop application for image content from the origin server during replay. These requests should not be captured.
If the previous criteria are met, then the image hit is dropped. When this feature is enabled, if
CaptureType=1 and the response content type is listed in the Include Extension
list, then both the request and the response are captured.
In the Statistics tab, the total count of dropped image hits is displayed in the
dropped due to businessIT mode and DelImages feature set metric.