Error: DynamoDB consistent read not working

A strongly consistent DynamoDB read must target a base table or local secondary index and explicitly request ConsistentRead=true. GSIs reject that setting. If the result still looks stale, verify the write completed and the read uses the same item, table, account, and intended Region.

• Await the write and check its response. A strong read cannot reveal a write that failed or has not completed.
• For GetItem, Query, or Scan, set the parameter on the actual request. A wrapper or cache can have separate consistency settings.
• On an MREC global table, strong reads reflect completed local writes and do not wait for remote replication. MRSC offers a different consistency model; verify the table’s mode.
• A strong Scan is not a snapshot of the whole table. Items can change between reads or pages, and another writer can modify an item immediately after any read.

Insufficient read capacity produces throttling; it does not silently downgrade a requested strong read. Check the exception and bypass an application cache when validating behavior. For invariants across writes, use conditions or transactions where supported rather than relying on a read-then-write sequence.

See global table consistency and AWS documentation.