restore checkpoint app image start slower when using sessionCache with infinispan

Restore eBuy checkpoint app image on OCP. The app is configured to have httpSessionCache with infinispan with the following in server.xml

<server>
    <featureManager>
	    <feature>appSecurity-4.0</feature>
	    <feature>sessionCache-1.0</feature> 
    </featureManager>

    <library id="DatagridLib">
	    <fileset dir="${server.config.dir}/datagrid" includes="*.jar"/>
    </library>

    <cacheManager id="CacheManager" uri="file:///${server.config.dir}/datagrid/infinispan_hotrod.props">
	    <properties infinispan.client.hotrod.uri="hotrod://${INFINISPAN_USER}:${INFINISPAN_PASS}@${INFINISPAN_HOST}:${INFINISPAN_PORT}"/>
        <cachingProvider providerClass="org.infinispan.jcache.remote.JCachingProvider" jCacheLibraryRef="DatagridLib" />
    </cacheManager>

    <httpSessionCache cacheManagerRef="CacheManager" cacheSeparator="_"/>

</server>

When restore eBuy app on OCP without using httpSessionCache, it takes less than 1 second or sometimes just a little over 1 second for the server to start, but when using httpSessionCache, it takes about 12 seconds for the first app pod to start and about 4 to 5 seconds for the later pods

I see that there are many org.infinispan.HOTROD messages in the logs that add more time to the server start time

pods' logs: pods-logs.zip

Slack discussion: https://ibm-cloud.slack.com/archives/C03MR7EC3NG/p1718041901887979?thread_ts=1717187652.885869&cid=C03MR7EC3NG and https://ibm-cloud.slack.com/archives/C03MR7EC3NG/p1718226721233629

Start with the server.xml configuration in the issue and compare the pod startup logs in pods-logs.zip with runs that omit httpSessionCache. Review the linked Slack discussions for existing investigation, then define done as identifying the source of the added startup delay and documenting or implementing a verified reduction.