KB485592: Additional jobs are spawned when executing Dossier using Connect Live dataset in MicroStrategy Library 2021 Update 4.1

Starting with the release of Strategy ONE (March 2024), dossiers are also known as dashboards.
SYMPTOM:
 
A large dossier using only Connect Live datasets spawn multiple jobs when executing from Library and never finish post upgrade to 2021 Update 4.1 from Strategy 2020 Update 1. (Beside Strategy upgrade, operating system migrates from Windows to Linux; and Metadata change from SQL Server to PostgreSQL)
The dossier is time consuming and the successful execution from web takes over twenty minutes. The Developer>Job monitor shows four jobs "Running dossier xxx" spawned and numerous child jobs. Total job number is quadrupling. The ID column behind the Description indicate it is the same dossier.

NOTE:
The symptom doesn’t expose in Web. And the total job number is expected to be equal to the number of visualizations contained in the dossier.
Refer to Knowledge Base Article below for understanding every visualization within the dossier will become a job if dossier is powered by Live Connected datasets:
KB442449: Dossier Performance: Connect Live performance considerations
It is noticed in Kernel XML API trace, Intelligence Server gets dossier execution request (ro) from Library 4 times, 1 request every minute (corresponding to the additional jobs spawned timestamp in the monitor) as shown below; tomcat access log trace show two requests (POST and Get) come from two different IP addresses
SERVER:CastorServer][PID:27505][THR:139872972478208][Kernel XML API][Trace][UID:2A8049C74BB89DEBC85EAF9BCBE9B8F9][SID:C4CF88B814DC23FF86F9CA2582E91F10][OID:0][XMLRequest.cpp:139] XML Command: <st><sst><st><cmd><ro sid="0156B3BB9F07E0279629ACAF296A84A7" pgd_login="EF1747E944EE2213CB6AB8A8D6A72CBF" apply_manipulation_mode="1" exef="49156"><doc><object_id>EEA050C447469F296127FF9D27FCE982< 
Network trace indicates the environment has configured awselb with 5 mins timeout setting.

Refer to Knowledge Base Article below on how to enable Network Trace with Chrome
KB438738: How to collect a HTTP request trace using Chrome or Internet Explorer 11 Developer Tools
CAUSE:
 
The proxy has retry logic on tcp connection errors and the retry timeout is 1 minute.
 

• In the web, when executing a large dossier that takes a long time to finish, the browser will keep refreshing and sending out requests every few seconds to check if the execution is done.

• While in the library, it's a different logic which sent request but wait for server's response. If the consuming time is long, it will hit ALB retry logic.
   

Thus, Strategy Web app that deployed to the same tomcat servers and is behind the same ALB did not exhibit the behavior. 
ACTION:
Adjust the setting in Load Balancer or proxy to prevent dossier execution request being intercepted.
 
WORKAROUND:
Deploy an env without Load Balancer awselb.
 
3rd party warning:
Third Party Software Installation WARNING
The third-party product(s) discussed in this technical note is manufactured by vendors independent of Strategy. Strategy makes no warranty, express, implied or otherwise, regarding this product, including its performance or reliability.