| Role | Primary Responsibilities |
|---|---|
| Users | - Interact with Table Storage through applications rather than direct access - Require assigned permissions for both user accounts and service principals |
| Administrators | - Also known as Database Administrators (DBAs) - Grant and revoke access rights to control who can perform specific actions - Perform backups and restores for recovery |
| Analysts & Developers | - Determine optimal PartitionKey and RowKey values to distribute data and speed up queries- Define entity properties to enforce a consistent schema - Monitor performance and refine queries |
Azure Table Storage supports a semi-structured schema: each entity can have different properties. However, aligning on a core set of attributes per row prevents uncontrolled schema growth and simplifies data handling.
Key Takeaways
- Deployment Options
- Native Azure Table Storage
- Azure Cosmos DB Table API for global distribution and low latency
- NoSQL Key–Value Model
- Each entity (row) holds all related data for a single transaction
- Required Keys
- Every row must include a
PartitionKey,RowKey, andTimestamp
- Every row must include a
- PartitionKey
- Determines how data is distributed across storage nodes
- Narrow your query to a single partition to limit scanned rows and improve performance
- RowKey
- Unique within its partition and automatically indexed for fast lookups
- Requires an exact
PartitionKeymatch to leverage the index
- Flexible Schema
- Store varying properties per row
- Maintain a baseline attribute set to ensure consistency and manageability
An unbalanced partition strategy (e.g., “hot” partitions) can cause throttling. Monitor access patterns and adjust your
PartitionKey design to distribute load evenly.