How to decide the primary key on CosmosDB ?

In Cosmos DB, the primary key is defined by the partition key and the unique ID (id) of each document:

• id: A unique identifier for each item in a container.
• partition key: Determines how data is distributed across partitions.

✅ How to Choose the Right Partition Key

Understand Your Access Patterns

• Choose a key that is frequently used in queries and filters.
  • Example: If you're building a marketplace like Trampo Square, and users often query by userId, then userId is a good candidate.
  Ensure High Cardinality

  - The partition key should have **many unique values** to avoid hot partitions.
  
     - Avoid keys like `country` or `status` if they have few distinct values.
  
     **Balance Read/Write Load**
  
        - Distribute traffic evenly across partitions.
  
           - If one key (e.g., `productId`) gets 90% of traffic, it will cause throttling.
  
           **Avoid Large Items in One Partition**
  
              - Cosmos DB limits partition size to **20 GB**.
  
                 - If all items with the same key exceed this, you’ll hit limits.
  

1. Consider Composite Keys (if needed)
   • You can simulate composite keys by concatenating values (e.g., userId|orderId).In Cosmos DB, the primary key is defined by the partition key and the unique ID (id) of each document:
   • id: A unique identifier for each item in a container.
   • partition key: Determines how data is distributed across partitions.
   ✅ How to Choose the Right Partition Key
   1. Understand Your Access Patterns
      • Choose a key that is frequently used in queries and filters.
      • Example: If you're building a marketplace like Trampo Square, and users often query by userId, then userId is a good candidate.
   2. Ensure High Cardinality
      • The partition key should have many unique values to avoid hot partitions.
      • Avoid keys like country or status if they have few distinct values.
   3. Balance Read/Write Load
      • Distribute traffic evenly across partitions.
      • If one key (e.g., productId) gets 90% of traffic, it will cause throttling.
   4. Avoid Large Items in One Partition
      • Cosmos DB limits partition size to 20 GB.
      • If all items with the same key exceed this, you’ll hit limits.
   5. Consider Composite Keys (if needed)
      • You can simulate composite keys by concatenating values (e.g., userId|orderId).