It's probably worth opening a case with support here. You'll want to look at the basic server sizing (v7.3 uses more memory the v7.2, so you may need to up the resource in the server / VM), which specific service (ie server, acq, elastic, mysql) is growing memory usage over time, as well as whether any of those processes are approaching the memory limits configured in the -xmx parameters in the registry.
I have seen other cases where tweaking the config to allocate more memory to the acq service ends up reducing the memory usage of elastic quite considerably, but I can't really say for sure whether that would fix your issue here.
It is a bit surprising that you would see memory utilization change as a result of a data source service pack - most of the memory OCI uses is controlled by registry / configuration file settings that are populated when you install OCI (this is why increasing memory on an OCI server may not have a major effect, as the OCI services will not get a lot of benefit from them). Because of the software stack OCI uses internally (Java, MySQL), memory management is a bit tricky - the OCI Installer attempts to set configuration settings to allow OCI to use as much memory as possible without causing any problems.
In general, however, OCI 7.3.0 needs more memory than earlier OCI versions. It may be that your environment needs > 32GB.
If your OCI server is sending OCI ASUP, PM me the site name, and I may be able to offer some suggestions