WebLogic Log Requests Test

This test monitors a web server access log and reports measures such as the number of requests that have been logged, the number of successful responses, the number of failed responses, etc., for every pattern that has been configured.

Target of the test : A WebLogic server

Agent deploying the test : An internal agent

Outputs of the test : One set of results for every search pattern being configured.

Configurable parameters for the test
Parameter Description

Test Period

How often should the test be executed.

Host

The host for which the test is to be configured.

Port

The port at which the application listens.

AbsoluteFileName

Specify the full path to the log file to be monitored.

RecordPattern

The records in the log file that need to be considered for monitoring will have to be provided in the RecordPattern text box. The pattern configuration should be in the following format: {f0}sep1{f1}sep2{f2}, where {f0}, {f1}, and {f2} represent the indexes of the first, second, and third fields (respectively) of the records logged in the log file, and sep1 and sep2 are the separators after {f0} and {f1} respectively. A separator can be a combination of any number of characters.

For example, take the case of a log file with the following entry:

192.168.10.7 - - [12/Nov/1998:09:40:40 -0500] "POST /soap/servlet/helloworld HTTP/1.1" 200 3834

To ensure that the above record is considered for monitoring, the record pattern will have to be specified as follows: {f0}- -{f1}"{f2}"{f3} {f4}, where {f0} represents the first field of the record, which is followed by the separator '- -', and so on.

SearchPattern

Of the records that match the configured RecordPattern, the eG agent will search for and monitor only those records which match the string patterns specified in the SearchPattern text box. To help you understand how to configure a SearchPattern, let us take the example of the following search pattern: IP1:ALL,F0:192.168.10.7*,F3: 200*,COUNT(*),AVG(F4).

  • Here, IP1 is just a display name that will be displayed in the eG monitor interface as a descriptor of this test.
  • The term ALL instructs the eG Enterprise system to consider only those records that fulfill all the conditions that follow. Alternatively, the key word Any can be used, which implies that the eG Enterprise system, while monitoring, will consider even those records that fulfill either of the conditions that follow. The conditions are:

    • F0: 192.168.10.7* indicates that for a record to be considered for monitoring, the first field (i.e. the field with index 0) of the record should begin with the IP 192.168.10.1. Alternatively, the condition can be configured as F0:192.168.10.7*+192.168.10.8*+192.168.10.9*, where '+' denotes an 'OR' operator. This configuration indicates that for a record to be considered for monitoring, the first field of the record should begin with any of the three values configured - i.e. 192.168.10.7, 192.168.10.8, or 192.168.10.9.
    • F3: 200* indicates that for a record to be considered for monitoring, the fourth field (i.e. the field with index 3) of the record should begin with the number 200. Alternatively, the condition can be configured as F3:200*+300*+400*, where '+' denotes an 'OR' operator. This configuration indicates that for a record to be considered for monitoring, the fourth field of the record should begin with any of the three values configured - i.e. 200, 300, or 400.
  • COUNT(*) returns the number of records that fulfill the configured criteria.
  • AVG(F4) returns the average of the values of all the fields with index 4 (i.e. the fifth field), in the records that match the configured criteria.

According to this specification, the eG Enterprise system, while taking a count and while calculating the average, will consider only those records where the first field starts with '192.168.10.1' and the fourth field starts with '200'. The number ‘200’ indicates a successful response. Therefore, this specification will report the metrics pertaining to only the successful responses for the IP patterns defined within the descriptor IP1 (i.e. 192.168.10.7*).

However, the test's configuration becomes complete only if the failure statistics are also extracted for IP1. Therefore, you will have to provide another search pattern for the descriptor IP1, so that the failure information is collected. The format of this pattern should be: IP1_FAIL: ALL,f0:192.168.10.7*,!f3:200*,COUNT(*),AVG(f4). Note that the descriptor names are the same, but the one meant for monitoring the failure cases, has been tagged as _FAIL. The specification !f3:200 indicates that the records with the number ‘200’ (in the fourth field) should NOT be considered for monitoring. ‘!’ is a NOT operator. Since ‘200’ represents a success state, !200 ensures that only the failed responses for IP1 are considered for monitoring.

The complete SearchPattern will hence be: IP1:ALL,f0:192.168.10.7*,f3:200*,COUNT(*),AVG(f4)#& IP1_FAIL:ALL,f0:192.168.10.7*,!f3:200*,COUNT(*),AVG(f4), where #& is the separator.

In the monitor interface however, the descriptor IP1 alone will appear, but when clicked, will display both the success and failure statistics for the pattern 192.168.10.7*. Therefore, it is imperative that the WLLogReqTest be configured in such a way that it tracks both the success and failure cases for every IP pattern configured for monitoring. Otherwise, the test will not function as desired. This implies that if an IP pattern IP2 is configured for monitoring successful responses, then an IP2_FAIL should follow to monitor the failed responses. Similarly, multiple patterns can be configured for monitoring, separated by ‘#&’.

Measurements made by the test
Measurement Description Measurement Unit Interpretation

Total requests

Indicates the number of account calls that are being made during a period of time.

Number

A high value of this measure indicates a heavy workload on the server.

Successes

Indicates the number of successful responses.

Number

Low value of this measure indicates less number of successful responses from the server.

Avg success bytes

Indicates the number of bytes of successful responses

Bytes

A high value of this measure indicates a high rate of successful responses.

Failures

Indicates the number of failed responses.

Number

 

Avg fail bytes

Indicates the number of bytes of failed responses.

Bytes

A high value of this measure indicates a high failure rate.

Avg bytes sent

Indicates the size (in bytes) of responses sent by the server.

Bytes

 

Note:

If any of the measures of this test returns the value -5, then such a measure will not be displayed in the monitor interface. On the other hand, if all the measures of this test return the value -5, then all the measures will appear in the monitor interface, but the value displayed for each measure will be "Not Available".