api | web | docs: back up Redis to S3 on a schedule

[yekta]

Redis was the only database without backups, so its data lived on a single volume with no off-site copy. Redis now uses the same backup bucket, schedule and retention settings as the other databases.

Resolves UNB-404.

• api: When S3 is enabled, the Redis definition renders a <name>-s3-backup CronJob. It dumps pod 0 (the primary in standalone and replication mode) with redis-cli --rdb and checks the dump with redis-check-rdb. It then uploads the dump as a timestamped .rdb file and deletes the oldest files beyond the retention count. Redis now imports the shared S3 schema, so the backup validation accepts it.
• web: The Backups section shows for every database type, both when creating a database and in its settings. The databaseTypesWithoutBackups list is removed.
• docs: Explain how to restore an RDB onto the Redis volume. The AOF conversion init container turns it back into an append-only file on start. Also note that the append-only file protects against a crash, not against losing the volume.
• The OpenAPI and generated client descriptions for s3_backup_bucket_id no longer say Redis is excluded.

Tests: added TestRedisDefinitionRendersS3Backups to check the rendered CronJob with backups on and off. Updated TestValidateBackupsSupported and TestEmbeddedDefinitionsSupportsBackups so Redis now supports backups.