Is DynamoDB eventually consistent?

By default, yes. DynamoDB reads are eventually consistent, meaning a read right after a write may briefly return stale data before replicas converge (usually within a second). You can request strongly consistent reads on tables and local secondary indexes by setting ConsistentRead to true; global secondary indexes are always eventually consistent.

Eventually consistent (the default)

An eventually consistent read may not reflect a just-completed write. Repeat the read a moment later and it returns the current value. This mode is the default and costs half of a strongly consistent read.

Strongly consistent reads

Set ConsistentRead: true to get the most recent committed data. This is supported on the base table and local secondary indexes (LSIs) — but not on global secondary indexes (GSIs) or DynamoDB Streams, which are always eventually consistent.

Why it works this way

DynamoDB replicates each write across multiple storage nodes. Eventual consistency lets reads hit any replica for speed and cost; strong consistency waits for the latest.

Go deeper

Read DynamoDB consistency for the full trade-offs. Download DynoTable to toggle read consistency on your queries.