I tried stress test with multiple connections.
When normally response time is 2 seconds, but 15 seconds response delay occured.
I confirmed to CPU usage and memory usage, but there is enough space.
I confirmed log, and
WARN [X.X] MonitorInvocationsTask delayed 15276ms.
was appeared in PegaCLUSTER.log.
In that time, PegaLog and access_log were not logged anything.
Why has this response delay occured?
Could you tell me possible solution or hypothesis if you have ?
please help me.
***Edited by Moderator: Pooja Gadige to add platform capability tag***
Are you doing stress test on your regular environment or on a copy?
is infra perfectly the same for servers and DB as well as their properties?
What kind of stress test are you doing? users activities, backend processing? is it a standalone application? or connecting other applications?
While reproducing your prod load, do you face the same delay?
There are lots of parameters to take into account and which explain this delay
Please provide some more info if you want some guidance
The infrastructure configuration is a simple configuration with a single node and a single DB.
Our stress test reproduces the user operation by Jmeter on regular environment(production environment),
but I have confirmed PegaCLUSTER.log that this event has occured by the regular user operating the case manager operation in the production environment about three times a month.
The User operations include listing by ReportDefiniton via Section, opening a case, and advancing the status using FlowAction. If you need any other information, please let me know.