MQ Monitoring agent: Storage and Memory Usage

Body

This blog entry has some considerations about the usage of storage and physical memory (RAM) by MQ monitoring agent.

Generally speaking the amount of memory used is based on the queue manager being monitored and the configuration.
In general, based on some customer configuration, the agent storage will grow and then stabilize after some time.
Then there will be some periods with large requests of storage to handle situation/workspace requests for data. 
This storage will be be released once the request data is returned to TEMS.

So in more detail, the amount of memory is based on a combination of configuration of queue manager and agent, as well as situation and workspace requests.
Specifically:

1) Sample collection

For monitored queues and channels, MQ monitoring agent has some attribute groups that are based on sampled collection.
This results in the last X samples being saved in memory. The number of samples is dictated by AGGRHIST. 
And each sample is taken every SAMPINT seconds. 
If you have AGGRHIST set to 15 and SAMPINT 300, so 15 * 5 = 1 hour and 15 minutes, so the memory should reach a steady state after about 1h and 15 minutes.
Note that there are some variations to this when history comes into play, since our agent will save history data that has not yet been requested if HISTORY(YES) is specified.

The amount of memory used will depend on the number of queues and channel definitions and active channel connections.  Data will generally stabilize after the 15 samples (or RETAINHIST value if  HISTORY(YES) is specified).

2) Situation/Workspace Requests

Memory is allocated to hold all the rows being returned for a situation or workspace request.  In addition to data allocated to return the data, there may also be some additional memory allocated while collecting this data.  This data will be requested and then released when data is returned to the TEMS.
The amount of memory needed for each will request will depend upon what data is requested, but it could be a significant amount in some cases.
For example, if many channel connections, Channel Status or Channel Statistics can return a lot of data, Queue Definitions and Queue Statistics can return a significant amount of data if a large number of queues are monitored, Queue Messages will return a lot of data if the queue has many messages.

3) MQI Statistics/Accounting Data

If the queue manager has MQI statistics and accounting turned on (STATMQI  ACCTMQI STATINT ACCTINT STATQ ACCTQ), then our agent will save in memory the most recent samples collected, based upon configuration specifications for RECENTACCOUNTINGSAMPLES and RECENTSTATISTICSSAMPLES.
The amount of memory needed for this is dependent upon QMGR specifications and application activity.  A large number of applications that connect and disconnect frequently and access a large number of queues could generate a lot of accounting data.
 

Subscribe and follow us for all the latest information directly on your social feeds: