Use this page to learn how to interpret performance data
for request metrics in trace record format.
The trace records for request metrics
data are output to two log files: the web server plug-in log file
and the application server log file. The default names for the log
files are
SystemOut.log and
http_plugin.log.
You might, however, specify these log file names and their locations.
The default directories for these log files are:
- plugin_install_root/logs/web_server_name/http_plugin.log
and install_root/profiles/profile_name/logs/server_name
and
In the WebSphere
® Application Server
log file the trace record format is:
PMRM0003I: parent:ver=n,ip=n.n.n.n,time=nnnnnnnnnn,pid=nnnn,reqid=nnnnnn,event=nnnn
-
current:ver=n,ip=n.n.n.n,time=nnnnnnnnnn,pid=nnnn,reqid=nnnnnn,event=nnnn
type=TTT detail=some_detail_information elapsed=nnnn
In the web server plug-in log file the trace record
format is:
PLUGIN:
parent:ver=n,ip=n.n.n.n,time=nnnnnnnnnn,pid=nnnn,reqid=nnnnnn,event=nnnn
- current:ver=n,ip=n.n.n.n,time=nnnnnnnnnn,pid=nnnn,reqid=nnnnnn,event=nnnn
type=TTT detail=some_detail_information elapsed=nnnn bytesIn=nnnn
bytesOut=nnnn
The trace record format
is composed of two correlators: a parent correlator and current correlator.
The parent correlator represents the upstream request and the current
correlator represents the current operation. If the parent and current
correlators are the same, then the record represents an operation
that occurs as it enters WebSphere Application Server.
To
correlate trace records for a particular request, collect records
with a message ID of PMRM0003I from the appropriate application server
log files and the PLUGIN trace record from the web
server plug-in log file. Records are correlated by matching current
correlators to parent correlators. You can create the logical tree
by connecting the current correlators of parent trace records to the
parent correlators of child records. This tree shows the progression
of the request across the server cluster. Refer to Why use request metrics? for an example
of the transaction flow.
The parent correlator is denoted by
the comma separating fields following the keyword, parent:. Likewise,
the current correlator is denoted by the comma separating fields following,
current:.
The fields of both parent and current correlators
are:
- ver:The version of the correlator. For
convenience, it is duplicated in both the parent and current correlators.
- ip:The IP address of the node of the application
server that generated the correlator. If the system has multiple IP
addresses, request metrics uses one of the IP addresses to identify
the system.
- pid:The process ID of the application
server that generated the correlator.
- time:The start time of the application
server process that generated the correlator.
- reqid:An ID that is assigned to the request
by request metrics, unique to the application server process.
- event:An event ID that is assigned to
differentiate the actual trace events.
Following the parent and current correlators, the metrics
data for timed operation are:
- type: A code that is represents the type of operation being timed.
Supported types include HTTP, URI, EJB, JDBC, JMS, COMMONJ_WORK_POOLED,
COMMONJ_TIMER, web services requester, and web services provider.
- detail: Identifies the name of the operation being timed (See
the following description of Universal Resource Identifier (URI),
HTTP, EJB, JDBCJ, JMS, asynchronous beans, and Web services.)
- elapsed: The measured elapsed time in <units> for this operation,
which includes all sub-operations called by this operation. The unit
of elapsed time is milliseconds.
- bytesIn: The number of bytes from the request that is received
by the web server plug-in.
- bytesOut: The number of bytes from the reply that is sent from
the web server plug-in to the client.
The type and detail fields that are described include:
- HTTP: The web server plug-in generates the trace record. The detail
is the name of the URI that is used to invoke the request.
- URI: The trace record is generated by a web component. The URI
is the name of the URI that is used to invoke the request.
- EJB: The fully qualified package and the method name of the enterprise
bean.
- JDBC The interface name and method name for that JDBC call.
- JMS: JMS includes the particulars of various JMS parameters
- Asynchronous beans: The detail specifies the name of the asynchronous
beans. Asynchronous beans include two types: COMMONJ_WORK_POOLED and
COMMONJ_TIMER.
- Web services: Web services include the particulars of various
web services parameters. Web services include two types: Web services
requestor and web services provider.
- SIB: Used for instrumentation in service integration bus including
message send/receive and mediation.
- JCA: J2EE Connector Architecture. The detail specifies the class
name in which the JCA call is made.
- JNDI: Used for JNDI naming look up. The detail specifies the JNDI
name.
- JMS send and receive: Generates the trace record by JMS sending
and receiving messages.
- SIB send and receive: Generates the trace record by SIB sending
and receiving messages.
When there are multiple servant regions
for an application server, there are multiple SystemOut.log files,
one for each servant region. Therefore, request metrics might log
the trace records in multiple SystemOut.log files.
The servant region that handles a request logs the relevant records
in itsSystemOut.log files. The pid in the current
request metrics correlator is the pid for the corresponding servant
region. If the system has multiple IP addresses, the IP in the correlator
could be one of them, but it should use the same IP for the same servant
region.
New feature: This topic
references one or more of the application server log files. Beginning
in WebSphere Application Server Version 8.0 you can configure the
server to use the High Performance Extensible Logging (HPEL) log and
trace infrastructure instead of using
SystemOut.log ,
SystemErr.log,
trace.log, and
activity.log files or native z/OS logging
facilities. If you are using HPEL, you can access all of your log
and trace information using the LogViewer command-line tool from your
server profile bin directory. See the information about using HPEL
to troubleshoot applications for more information on using HPEL.
newfeat