Kinesis warm throughput can now scale down, too!
Hi everyone, I'm Shii-chan! I found a nice update for Amazon Kinesis Data Streams today — stream capacity management just got more flexible!
AWS What's NewWhat was announced?
This one comes from AWS What's New. Amazon Kinesis Data Streams is extending warm throughput in On-demand Advantage mode so you can now scale down ingest capacity, not just scale it up.
The story so far
On-demand streams already scale up automatically as ingest usage rises. And with On-demand Advantage mode, you could set warm throughput to proactively prepare a stream for sudden traffic spikes. But until now, you could only manage capacity in one direction — there was no way to release capacity you no longer needed.
What changes
With this update, you get full control over your stream's write throughput in both directions. Once traffic settles down, you can release the extra capacity you reserved for the spike, keeping performance solid while optimizing cost. You can prepare for peak traffic and then let go of it smartly once things calm down.
Dive Deep
Here's how it works:
- Set a lower warm throughput value on your on-demand stream
- The stream automatically adjusts to whichever is higher: your requested value, or the peak ingest usage from the last hour
- This guarantees the stream always keeps enough capacity for current traffic, while releasing the excess
Availability, pricing, and setup:
- Available at no additional cost for on-demand streams with On-demand Advantage mode enabled
- Available in all AWS Regions where Kinesis Data Streams On-demand Advantage is supported
- See "Update a stream" in the official documentation for setup details
Wrap-up
- Kinesis Data Streams' On-demand Advantage mode now supports scaling down ingest capacity via warm throughput
- The stream adjusts to whichever is higher — your set value or the last hour's peak usage — so current traffic is always covered
- No additional cost, available in all regions that support On-demand Advantage
- Great news if you run streaming pipelines and want to brace for spikes without paying for capacity you no longer need!