We have an SLA agent processing multiple flows(we have to process large size of soap responses), while processing the agent requestor is using all the heap space and giving PEGA0028 Alert in logs, as this is SLA agent, I am not able to check the clipboard size from Requestor management in SMA, i have tried profiling agent requestor but it does not give clipboard size. So I want to know memory usage of this agent processing at multiple points in flows
please suggest below points
I) is there any java method or pega functions to get clipboard size so that i log this
ii) is there any Performance Profiler tools to get Clipboard size of the agent requestor processing
iii) and how to stop processing when the requestor is consuming all JVM heap memory
iv) what is a maximum size of a clipboard can a requestor have.
You can see the Clipboard size for requestors in SMA (Requestor Management page > select Requestor Name > click the Clipboard Size button). You can also see the Clipboard size using the Performance tool. I am not sure about modifying the size.
I could only find this discussion should be helpful beside that I don’t believe there is any other way.
I understand that from SMA and clipboard analyzer, we can see user requestor/clipboard size. Besides that,
1. Do we have configuration setting to set memory threshold limit on requestor level ? If is, once the requestor limit is reached, what the system will behave ?
2. For #2, will AES or other tool(s) have the feature to identify the session memory threshold is reached and alert the user ?
We want a tactical approach to limit the memory on requestor level so system will not run out of memory if one user accidentally pulled 500 MB data into clipboard. Also production support team don't want to monitor SMA 24/7 to get user clipboard sizing information, looking for an agent that can alert production support team if any user requestor have memory over threshold
There is no alert or limit on clipboard size.
There are already alerts in place for
- too many rows in a database query
- too many members on a page list
The alerts will be sent to AES. At present, AES does not consider these alerts to be 'critical' but you can probably override these rules through aescriteria ruleset.
It is possible to "automate SMA". Obviously, one can "web scrape" prsysmgmt, but better, the actual JMX Mbean operation to get requestor list with clipboard estimate is documented and may be called via any third party JMX tool. There's also an activity and service in PegaAESRemote ruleset that generically expose all the JMX Mbean operations to external monitoring tools (including AES). Would not be hard to write an agent activity on AES to poll all servers periodically to check overall clipboard page size for each user.