The trace based performance analysis collects the trace information for the functionality layer.

The following table describes the information output to trace based performance analysis file (CSV format) from the performance tracer. The collection items differ in CTM and the other function layers. Note that the output items are different depending on each collection point such as the existence of the additional information. For details on the items output for each collection point, see the chapter 8. Trace Collection Points and PRF Trace Collection Levels of the Trace Based Performance Analysis.

Table 7-4 Information output to the trace based performance analysis file (for the trace based performance analysis)

Trace information header	Description	Range of values
PRF	Record status of that process (normal and abnormal).	Either of the following is output: Normal: Rec Abnormal: ErrRec
Process	Process ID of the process that acquired the trace information.	A decimal number of ten digits is output.
Thread	Thread ID and hash value of the thread in the process that acquired the trace information#1.	Thread ID: A decimal number up to twenty digits is output. Hash value: A decimal number up to ten digits is output.
Trace	Trace sequence number of the thread in the process that acquired the trace information.	A decimal number of ten digits is output.
ProcessName	Process name	A string#2 displaying the process up to 32 characters is output.
Event	Event ID showing the trace collection point.	A hexadecimal number of six digits (include 0x also in six digits) is output#3.
Date	Date when the trace information is acquired.	The date is output in the yyyy/mm/dd format. yyyy: Year mm: Month dd: Day
Time	Time when the trace information is acquired (Hours: Minutes: Seconds).	The time is output in the hh:mm:ss format. hh: Hours mm: Minutes ss: Seconds
Time(msec/usec/nsec)	Time when the trace information is acquired (milli seconds/micro seconds/nanoseconds).	The time is output in the ms/us/ns format. ms: Milli seconds us: micro seconds ns: nanoseconds
Rc	Return code.	A hexadecimal number of sixteen digits (include "0x" also in sixteen digits) is output. Normal:0 Abnormal:1 (or other than 0)
ClientAP IP#4	IP address of the client application that acquired the trace information.	The IP address is output in the aaa.bbb.ccc.ddd format.
ClientAP PID#4	Process ID of the client application that acquired the trace information.	A decimal number of ten digits is output.
ClientAP CommNo.#4	Communication number of the client application that acquired the trace information.	A hexadecimal number of 18 digits (include 0x also in 18 digits) is output.
RootAP IP#5	IP address of a root application that acquired the trace information.	The IP address is output in the aaa.bbb.ccc.ddd format.
RootAP PID#5	Process ID of a root application that acquired the trace information.	A decimal number of ten digits is output.
RootAP CommNo.#5	The communication number of the root application that acquired the trace information.	A hexadecimal number of eighteen digits (include "0x" also in eighteen digits) is output.
SendSCD IP#5	IP address of the request source CTM.	The IP address is output in the aaa.bbb.ccc.ddd format.
SendSCD PID#6	Process ID of the request source CTM.	A decimal number of ten digits is output.
ReceiveSCD IP#6	IP address of the request destination CTM	The IP address is output in the aaa.bbb.ccc.ddd format.
ReceiveSCD PID#6	Process ID of the request destination CTM.	The decimal number of ten digits is output.
INT	Interface name for each collection point.	A string#7 up to 33 characters is output.
OPR	Operation information related to the collection point.	A string#7 up to 33 characters is output.
LookupName#6	Lookup name.	A string#7 up to 33 characters is output.
OPT#8	Additional information for each collection point.	A hexadecimal number string up to 514 characters is output.
ASCII	ASCII character output of the additional information for each collection point.	The OPT contents is output as ASCII character strings within 514 characters.

#1

There are cases when the hash value of a thread is not output in the trace information acquired by the redirector and CTM.

#2

The process name is decided as described below:

• In the case of EJB client applications
  The name specified in the system property ejbserver.server.prf.processName of the EJB client application.
  If this system property is not specified or null character is specified in this property, the process name will be EJBClient.
  
• In the case of J2EE server, batch server, or Web container server
  The server name will be the process name.
  
• In the case of redirector
  If Hitachi Web Server is used as the Web server, "RD-Web-server-waits-to-receive-request-on-this-port-number" will be the process name. If Microsoft IIS is used, the process name will be Redirector.
  
• In case of CTM
  It is the name of each CTM process.
  

#3

An event ID is allocated to each trace collection point of the function layer. For details, see 8. Trace Collection Points and PRF Trace Collection Levels of the Trace based performance analysis.

#4

This is a component of client application information.

#5

This is a component of root application information. At the trace point that is output to the Web container, there are cases when the IP address, process ID, and communication number of the root application that acquired the trace information are output as 0.0.0. 0/0/0x0000000000000000. For details, see 7.7.7 Investigating the Log Using the Root Application Information.

Note that when you operate an HTTP server in a dual stack environment of IPv4 and IPv6, the IPv4 address is output to the process ID of the root application. Also, if the client address is IPv6 and is 33 characters or more, edit and output the trace information event IDs 0x8000 and 0x8100.

• For event ID 0x8000: 32-characters-from-the-beginning + *
• For event ID 0x8100: 16-characters-from-the-beginning + * + 16-characters-from-behind

In such cases, if both event IDs 0x8000 and 0x8100 are referenced, you can obtain the entire client address.

#6

The information that is output only for CTM. "****" is displayed in the layer other than CTM. In the systems executing batch application, the schedule group name is displayed.

#7

If the interface name, operation name, and lookup name exceeds 32 characters, the name is changed to 33 characters and is output with one of the following methods. For details, see 8. Trace Collection Points and PRF Trace Collection Levels of the Trace based performance analysis.

First 32 characters + *

First 16 characters + * + last 16 characters

* + last 32 characters

#8

Some function layers have trace collection points that output the entry time to OPT. The entry time is the time at which the entry trace corresponding to the trace of the trace collection point is output. For example, in the case of trace collection point of a redirector, the entry time output at the trace collection point when the process of sending HTTP response body information to a Web server finishes (0x8104), will be the time to start sending the HTTP response body information from a Web container (0x8004).

Note that the entry time is output in 16 bytes that is the time elapsed from 01/01/1970 00:00:00. The first 8 bytes of the value are seconds and the last 8 bytes of the value are microseconds. However, in the case of trace based performance analysis of redirector, the last 8 bytes are milli seconds.