The best method of improving the performance of a 2018.11 UCMDB is to upgrade to at least 2022.11. CMS 2023.05 is better.
There were a lot of fundamental changes which improve performance.

what is the roadblock for your upgrade?

Upgrading will take some time for us.

We need to keep it working in the meantime but its performance is highly degraded.

Then it depends on what are the signs of bad performance. 
Is the DB slow?
High number of CIs, duplicates?
Aging is disabled?
Bad Discovery/integration schedule with a lot of overlapping?

If you are asking for improvements then you need to know first what's the bottleneck.

We are checking with the DB team.

There are high number of CIs.

Aging is also disabled.

Aging will have the side effect of accumulated CIs which no longer have a correspondent in the real topology so they should be deleted. It will be tricky to determine what aged CIs need to be manually deleted.
In 2019 we implemented Nested Classes which improves a lot the server performance. installed_SW, process and and windows_service. This is just one of the changes implemented to support larger, faster or containerized UCMDB instances.

In your scenario there is no magic trick to improved drastically the performance. As I saw in a lot of similar scenarios, most likely you will lose more time on debugging and troubleshooting then the time lost on an upgrade.