AWS Updates 2026-07-16 | Cloud Provider News
Store logs more cheaply with automatic tiering
CloudWatch Logs now has intelligent storage tiering. It automatically moves logs between Standard, Infrequent Access, and Archive Instant Access tiers based on access patterns.
For teams with large log volumes, this can simplify retention strategy quite a bit. It reduces the manual effort of deciding where each log set should live and helps keep storage spend under control.
Better control your storage spend in Amazon MQ
Amazon MQ just added a more flexible way to size RabbitMQ storage. You can now configure EBS storage size independently of the RabbitMQ broker instance type, which gives you a cleaner way to right-size storage without changing compute.
In practice, this should make storage planning a lot less rigid for RabbitMQ workloads. Instead of overprovisioning just to get the disk size you want, you can align storage more closely with actual usage.
Tune RDS storage more often without waiting around
Amazon RDS now allows up to four storage modifications in a rolling 24-hour window. That’s a bump from the previous limit and gives teams more room to adjust storage as demand changes.
For FinOps teams, that means more agility in storage right-sizing. It should be easier to react to workload changes without turning every capacity shift into a bigger, more expensive decision than it needs to be.
Recover faster with AWS-to-AWS disaster recovery
AWS Elastic Disaster Recovery got a faster path for source servers already running on AWS. AWS says it can skip unnecessary preparation steps and reduce recovery time by up to 65% for Windows and up to 40% for Linux.
Make failovers more predictable with faster EBS hydration
AWS Elastic Disaster Recovery also added EBS volume initialization rate support. This helps recovered volumes reach full performance faster during drills and recoveries.
For FinOps, that can lower the hidden cost of underperforming recovery environments. It gives teams a little more confidence that recovery storage will behave the way they expect when it matters most.
Protect AI workloads from spend-heavy attacks
Amazon GuardDuty now includes AI Protection for AWS AI workloads. It’s designed to detect threats aimed at AI services like Bedrock and SageMaker, including cost-harvesting attacks that try to drive up GPU and token usage.
That’s a direct FinOps safeguard because attack traffic can turn into unplanned AI spend very quickly. Detecting those patterns early helps teams avoid paying for malicious or wasteful consumption.
Cut Lambda storage friction with self-managed S3 code
AWS Lambda now supports referencing function code directly from customer S3 buckets. That means you no longer need a Lambda-managed copy in the same way as before, and AWS also notes an increased managed quota.
The clear benefit is less quota pressure and less operational overhead. If you were running into the previous 75 GB limit, this change gives you another way to keep deployments moving without juggling storage workarounds.
It also centralizes storage cost in S3, which can make cost management cleaner. For teams handling lots of function artifacts, that’s a simpler model to operate and finance.
