The following lists my recommendations for monitoring WebSphere Message Broker or IBM Integration Bus with OMEGAMON XE for Messaging:

• For monitoring WebSphere Message Broker v7 and v8, and IBM Integration Bus v9, use v7.1.0 of the WebSphere Message Broker monitoring agent.   That is FMID HKQI710 on z/OS.  You'll need the latest PTF levels for later broker releases, which includes PTFs UA71110 and UA69753.  On distributed platforms for the same agent in ITCAM for Applications, this level is equivalent to the v7.1.0 Fixpack 1 + Ifix 1 level; v7.1.0 FP1 can be downloaded from Passport Advantage with part number CIJM0EN, and then the Ifix 1 that goes on top is at: http://www-01.ibm.com/support/docview.wss?uid=swg24035846.  

• Since queue monitoring is also usually necessary for broker environments, it is recommended to use the WebSphere MQ monitoring agent along side the WebSphere Message Broker monitoring agent so that you can automatically detect queues getting full, queues with messages sitting on them for too long, channels not running, and other problematic situations.   With the WebSphere MQ monitoring agent, there is not much that you must enable externally for recommended monitoring, except that if you are interested in Oldest Message Age and Queue timing (latency) attributes for monitoring your queues, you must turn on Queue Monitoring (MONQ) for queues in your queue manager (at least for the ones that are involved with your broker flows).  

• To see all the data produced for the broker, refer to the attribute groups applicable to broker v7 and later versions, which are documented here:   http://publib.boulder.ibm.com/infocenter/tivihelp/v15r1/topic/com.ibm.omegamon.mes_doc_7.1/r-attr_broker7.html. Various attribute groups are referenced in items below.                                                                                                                               

• Archive Accounting Statistics will provide the message flow counts and rates of messages, elapsed and CPU timings, sizes of messages, and error counts.   Archive Accounting Statistics generally are preferred to using CandleMonitor Node statistics nowadays.  They are easier to enable, do not require changes to message flows like CandleMonitor node usage does, and they contain CPU timings. 

The following attribute groups provide these statistics at four levels:

Accounting Message Flow Statistics attributes

Accounting Node Statistics attributes

Accounting Terminal Statistics attributes

Accounting Thread Statistics attributes

You enable the data in the broker by using the mqsichangeflowstats command.  It is available as a take action command from the WebSphere Message Broker monitoring agent as well.  To enable the recommended "archive" accounting statistics across the broker, with all levels of data available, issue the following command: 

z/OS:             F <broker>,changeflowstats  --a=yes -g=yes -j=yes -c=active -n=advanced -t=basic -o=xml 

distributed:      mqsichangeflowstats <brokername> -a -g -j -c active -n advanced -t basic -o xml

Notes:   
(a) -o xml setting is required for the monitoring agent to get the data. The agent will automatically start receiving the data with no other action required.
(b) It is not recommended to use the "snapshot" (-s option) accounting statistics in production monitoring, except if required for debugging an issue with a particular message flow, because snapshot generally produces too much data.  The archive and snapshot data is all exactly the same, just the interval for snapshot record production is very short, so it can impact performance and agent memory requirements. 
(c) You can modify the interval for archive accounting statistics by using the mqsichangebroker command (also available as a take action command of the monitoring agent).  You will likely want to do this because the default is 60 minutes.  This is the command to issue to specify a 5 minute interval (minimum is 1 minute):

z/OS:              F <broker>,changebroker  -v=5  

distributed:    mqsichangebroker <brokername> -v 5   

• Resource Statistics of six types are available via the following attribute groups:

JVM Resource Statistics attributes

JDBC Connection Pools Resource Statistics attributes

ODBC Resource Statistics attributes

File Resource Statistics attributes

Parsers Resource Statistics attributes

SOAP Input Resource Statistics attributes

You enable the data in the broker with the mqsichangeresourcestats command.  The command to enable for all execution groups in the broker is as follows:

z/OS:              Run command "mqsichangeresourcestats <brokername> -c active" by customizing and submitting BIPCHRS  

distributed:   mqsichangeresourcestats <brokername> -c active 

If you want to enable only for a particular execution group, add  "-e <egname>" to the command. 

• To have node detail data available for message flow topology and processing nodes workspaces, an agent parameter requires a change.  This parameter affects data availability for the the following attribute groups:

Message Processing Nodes attributes

Processing Node Attributes

Message Flow Node Topology attributes

 For this data, set the following parameter to YES in your agent parameter XML file:

defaultCollectNodeData="YES"

The parameter file is documented here: http://publib.boulder.ibm.com/infocenter/tivihelp/v15r1/topic/com.ibm.omegamon.mes_doc_7.1/c-para_prfrmn_intro.html.

The reason why this data is not enabled by default is that the amount of this data can be quite high, depending on the number of message flows, nodes, and attributes in your broker, and collecting this data can use more resources than you desire for your agent.  If having the Message Flow Topology and node detail data available in the TEP is interesting to you, it is recommended that you try enabling the data, and then check to make sure your agent still performs as desired.

• The attribute groups most recommended for historical collection are given here.  However, most attribute groups are eligible for enabling for historical collection and data warehousing.  Once in the TDW, a Cognos data model is provided for Tivoli Common Reporting with some sample reports.

Accounting Message Flow Statistics attributes

Broker Status attributes

Execution Group Status attributes

Message Flow Status attributes (if you want the status of each message flow recorded historically)

Broker Status Events attributes (only if you are interested in configuration events being recorded historically to tell when the broker was deployed with a change)

• All attribute groups that begin with the words "Monitor Node" will not have data if the CandleMonitor node is not configured and deployed in your message flows.  As mentioned above, Accounting Statistics are now usually preferred because of easier enablement and CPU timings being available.  The CandleMonitor node dates back to broker releases in which there were no other statistics available.

Tags:  best_practice omegamon broker mq zos