My team maintains a moderately sized redshift instance as a query source for analytics and we are looking for best practices in keeping a easy to refresh near production instance available with periodic refreshes to match our sprint cadence.
Our concerns are:
1. maximum automation, minimal manual effort
2. minimize any effects on production
3. maintain preproduction specific users/privileges/groups and passwords
4. recreate any work in-flight on the preproduction instance (only included for completeness, we know this is code management)
5. minimize downtime on preproduction
6. Raw data is maintained using Spectrum and may need to be re-configured
Simple redshift snaps satisfy all but the security requirements Table based unload/loads are time consuming and carry the risk of introducing referential integrity issues
Aucun commentaire:
Enregistrer un commentaire