Consider switching Grafana from SQLITE to PostgreSQL and move initial config to provisioning files

[DavidEdell]

We are currently using Grafana's default SQLITE3 DB, which we are including in our repository for initial configuration. We do not currenly have any backup of our default Grafana configurations outside of this binary database file.

We should consider migrating Grafana to use the PostgreSQL database server (new DB, same instance) we are already running within ANMS.

This has performance and maintainability benefits in the long-term, particularly for more complex use-cases in future production deployments of ANMS. Additionally, storing the binary sqlite database in version control is less than ideal - we should store a .sql file backup if we are unable to migrate initial configuration setup to configuration files.

[DavidEdell]

Most, if not all, of our grafana configuration should be definable in configuration files, which we are already using for datasources. See https://grafana.com/docs/grafana/latest/administration/provisioning/ or https://grafana.com/docs/grafana/latest/as-code/

Items we should migrate:

• plugins - Should be defined in provisioning directory instead of directly mapped, though we do not currently have any plugins in the master branch.
• dashboards - Reminder: Click the 'share' icon in the Grafana interface to export existing dashboards and use the 'for sharing externally' option.
• .gfenv - Should be moved into docker-compose.yml and/or .env to consolidate settings.

After making these changes we no longer require the grafana_vol folder.

[d-linko]

handled by #300