I faced this issue after migrating from Magento 1.9 to 2.3.5 and managed to solve it.
The issue is caused by the reindexing process not being able to finish it's task. So basically when ever you run
php bin/magento cron:run or
php bin/magento indexer:reindex you will get an additional temp table which is not being cleaned up.
In my case, the issues was caused by the CLI php version. I used php 7.3 when calling cron, but this raised some errors during the reindex process, so I switched to using php 7.1 when calling
For Ubuntu, you can try using
/usr/bin/php7.1 bin/magento cron:run.
The correct php version will vary depending on your Magento 2 version (older Magento versions will require older php versions for running cron jobs).
If you don't have access to your hosting server via SSH, you will need to ask your hosting provider to set this up.