The reality is, most running systems will still need specialists to manage the following:-
- Backup and Recovery: Although hugely simplified, it’s important to be able to quickly recover from system or human error, and quickly restore corrupted data.
- Manage Costs: Although it’s possible to spin up, resize or suspend a compute clusters within seconds, somebody must maintain control, and be accountable for costs. As billing is charged by the second, it becomes increasingly important to suspend warehouses when they are no longer needed, and while there’s the temptation to turn the Data Lake into an Ocean, somebody needs to organise the data, and ensure resources are the money is spent wisely.
- Software Development Lifecycle Management: Will always be a challenge for multi-terabyte systems. As compute power becomes increasingly available, business users will want to do more, but somebody will still need carefully manage code and data migration from development, test and production.
- Tune Warehouses: Snowflake gives the unique opportunity to adjust both the size of the cluster (to run faster, more complex queries), and the concurrency which means it can be adjusted to handle more users. The DBA needs to monitor the system, and adjust virtual warehouses when appropriate.
In the next article, I’ll explain in more detail which tasks still remain for DBAs and System Administrators. It’s fair to say however, with the wholesale migration to The Cloud, the landscape is changing, and DBAs will be freed up from the more mundane tasks.