Follow the following list:
- Use automated tasks to regularly restart system components. Do archival and backup tasks and other regular resource-consuming tasks such as GWC on low-use hours.
- When possible, in combination with clustering for high availability, enable logging and monitoring, making sure these processes are limited to only what's needed, for example, using the production logging profile and doing audit monitoring to a file rather than a database.
- Use an external data directory and make sure that this is also backed up in an automated routine fashion, and do the same with database backups.
