关于Azure Mobile Apps Easy Tables的限制与重要考量的技术咨询

Hey there, let's break down the key limits and critical considerations for Easy Tables in Azure Mobile Apps—this info is essential to avoid headaches when building and scaling your app:

Row Size Limit #

• The hard limit for a single Easy Table record is 64KB. This refers to the size of the serialized JSON payload of the row—so long text fields, nested objects, or large binary data will eat into this quota. If you exceed this, write/update operations will fail outright.
• For large data (like images, documents, or long-form content), don’t store it directly in the table. Instead, upload it to Azure Blob Storage and store only the blob’s reference URL in your Easy Table row.

Row Count Limit #

• Easy Tables are built on top of Azure SQL Database (or SQL Server), so the actual row count limit depends on your SQL tier’s storage capacity. For example, the free SQL DB tier has 1GB of storage—your max rows will be determined by average row size. Paid tiers scale with storage and compute (DTUs/vCores), so they can support massive datasets in theory.
• The bigger gotcha here is query performance. Once your table passes ~100k rows, unindexed queries (especially filters or pagination) will get slow. Always add indexes to fields you regularly query on to keep performance snappy.

• Permissions & Security: By default, Easy Tables are set to "Anybody with the application key can access"—this is fine for testing, but never leave this enabled in production. Lock it down to only authenticated users, or build custom APIs for granular access control (e.g., users can only view their own data). Also, never hardcode the application key in client-side code—use Azure AD or other robust auth methods instead.
• API Rate Limits: Azure Mobile Apps enforces rate limits on API requests. The free tier allows 5,000 requests per day, while paid tiers have higher quotas based on your pricing plan. If you’re building a high-traffic app, map out your expected request volume ahead of time to avoid unexpected throttling.
• Transactions & Consistency: Easy Tables don’t support cross-table transactions. If you need atomic operations across multiple tables, you’ll have to implement this via custom APIs or SQL stored procedures. Also, note that Easy Tables use eventual consistency by default—plan accordingly for use cases that require strong consistency.
• Data Type Restrictions: Not all SQL data types are supported. Avoid types like timestamp, sql_variant, or hierarchyid—stick to supported types like int, nvarchar, datetimeoffset, bit, and float. For complex objects, serialize them to JSON and store them in an nvarchar field.
• Offline Sync Limits: If you’re using offline sync, keep in mind that the local SQLite database has its own constraints (though the 64KB row size limit from Easy Tables still applies during sync). The default sync batch size is 50 records—you can adjust this, but larger batches may slow down sync times, especially on poor network connections.

内容的提问来源于stack exchange，提问作者dev-cc