A game studio's DynamoDB table uses auto scaling, reaches the maximum capacity on that policy during the evening peak, and still returns throttle errors. What should they change?
Select an answer to reveal the explanation.
Short Explanation
The elevator is already pressed against the top floor you set, and mashing the button will not build a new floor. You maxed the DynamoDB auto scaling ceiling, so the evening peak throttles until that maximum is raised. Scaling earlier, or caching reads, does not lift the same ceiling.
Full Explanation
DynamoDB auto scaling will not set provisioned capacity above the maximum on the scalable target. Once the table is at that ceiling and still throttling, the policy is doing what it was configured to do, and the maximum has to be raised before the evening peak can grow. That maximum is a hard stop on the scalable target, not a hint the service can exceed when throttles show up. Read throttles on the base table, not only on a quiet index. Lowering the target utilization only changes how full the table is before a scale-out starts, and the scale-out still stops at the same maximum. A global table replicates items to another Region, while writes sent to this table still consume this table's capacity. DAX caches item reads and does not accept the writes that are being throttled. Caveat: a higher maximum raises the capacity the table is allowed to reach, and a single hot key can still throttle after the table ceiling is lifted. Operational check: during the next peak, confirm consumed capacity rises above the old maximum and the throttle count drops.