Additional DynamoDB Features and Concepts

Data Model and Keys

• Partition Key (Hash Key): Required for all items, must be unique if no sort key

• Sort Key (Range Key): Optional, enables composite keys and more complex queries

• Composite Key: Combination of partition key and sort key

• Item Size: Maximum size of 400 KB including all attributes

Secondary Indexes in Detail

Local Secondary Index (LSI):

• Must be created at table creation time

• Same partition key as base table but different sort key

• Limited to 5 LSIs per table

• Shares provisioned throughput with main table

Global Secondary Index (GSI):

• Can be created/deleted anytime

• Different partition key and/or sort key from base table

• Limited to 20 GSIs per table

• Has its own provisioned throughput

Data Types

• Scalar Types: Number, String, Binary, Boolean, Null

• Document Types: List, Map

• Set Types: Number Set, String Set, Binary Set

Time To Live (TTL)

• Automatically delete items after an expiry timestamp

• Useful for session data, temporary data, logs

• No additional cost for TTL deletions

• TTL deletions can be captured in DynamoDB Streams

Error Handling

• ProvisionedThroughputExceededException: When requests exceed provisioned capacity

• ThrottlingException: When too many control plane operations

• ResourceNotFoundException: When accessing nonexistent table

• Exponential backoff for retry operations

Monitoring and Metrics

• CloudWatch Metrics:

  • ConsumedReadCapacityUnits

  • ConsumedWriteCapacityUnits

  • ReadThrottleEvents

  • WriteThrottleEvents

  • ThrottledRequests

  • SystemErrors

Batch Operations

• BatchGetItem: Retrieve up to 100 items from multiple tables

• BatchWriteItem: Put or delete up to 25 items in multiple tables

• Automatic retry for unprocessed items

Query and Scan Operations

Query:

• More efficient than Scan

• Requires partition key value

• Optional sort key conditions

• Can use any LSI or GSI

• Forward and backward scanning

• Limit parameter to control items per page

Scan:

• Examines every item in the table

• Optional filter expressions

• Parallel scan capability

• Higher resource usage

• Pagination support

Best Practices

• Use appropriate partition keys to distribute data evenly

• Implement retry logic with exponential backoff

• Use batch operations when possible

• Choose between provisioned vs on-demand based on workload predictability

• Use sparse indexes to minimize storage and write costs

• Compress large attribute values

• Use projections in secondary indexes to minimize data transfer

Limits and Quotas

• Maximum number of tables per region: 2500

• Maximum number of concurrent table operations: 500

• Maximum partition key length: 2048 bytes

• Maximum sort key length: 1024 bytes

• Maximum number of projected attributes in LSI: 20

• Maximum number of attributes per item: No fixed limit