The ThingWorx database stores the configuration data for ThingWorx and Navigate, users and groups, the Navigate extensions and related data such as mashups and customizations. One of the most important things is the ThingWorx encryption keystore and password, which are necessary to access anything in ThingWorx that uses a password. If you lose those, that instance of ThingWorx fails.
No Windchill data is permanently stored on the ThingWorx database. Some minimal temporary information is probably stored so the “RECENT” tab in the OOTB apps can be populated, but this changes with use.
The ThingWorx database is relatively small, and so many customers choose to install it on the same server with ThingWorx. On my own installation of Navigate 1.8 (PostgreSQL 9.4) that has been up over a year, the PostgreSQL folder (including the software as well as the data) takes up 988 MB. On my Navigate 8.5.0 using PostgreSQL 10 it is 650 MB.
Another consideration is performance. In extreme cases where ThingWorx and the database are installed on separate servers, in separate datacenters, at opposite ends of the country, Internet lag sometimes makes performance atrocious. Being in the same datacenter greatly improves that issue, and being on the same server may provide a small improvement over that. The same goes for where ThingWorx is installed in relation to Windchill. Cross-country datacenters have caused some serious lag issues.
Backup frequency is really a matter for your Disaster Recovery policy and risk management. While you don’t have any “business data” stored in the ThingWorx DB (its all in Windchill), how long it would take to restore the DB and update anything since the last backup would determine (in part) how long the system was down for users. The most critical information for that purpose is the encryption keystore/password. If you don’t have a backup of those that is current and reliable, you will most likely be doing a re-install from scratch if you lose the database. Customizations in particular could take a substantial amount of time to redevelop.
A typical corporate policy for a Production environment is a weekly full backup with daily incremental backups, but policies vary. It’s a balance between risk and cost, and only you can evaluate that.
