AWS Updates 2026-07-30 | Cloud Provider News
Organizations can now commit to Serverless Reservations with an all upfront payment option.
Redshift Serverless added an all upfront option for 3-year reservations
AWS says this can deliver up to 50% savings over on-demand rates. For teams with steady Redshift Serverless usage, that gives you a more predictable way to lock in lower RPU costs.
AWS Clean Rooms added bigger worker types for SQL analyses
SQL analyses can now run on 32 vCPU and 244 GB worker types.
Those larger workers are meant to speed up compute- and memory-intensive partner analyses, including complex joins and feature engineering. If you’re working through heavy jobs, faster completion can also help reduce overall compute spend.
Amazon Kinesis Data Streams added warm throughput scale-down
That gives fluctuating workloads more room to scale down when demand drops while still keeping enough capacity ready. It’s a straightforward way to improve cost control without giving up resilience.
Amazon EKS Provisioned Control Plane made HPA-driven scaling faster
Horizontal Pod Autoscaler sync concurrency is now up to 40 times the default.
That improves how quickly large clusters react to load spikes. In FinOps terms, faster scaling can help avoid both late capacity and the overspend that comes from overcompensating for slow scaling.
AWS also introduced a single log group for Bedrock AgentCore observability
Agent traces, prompts, and logs now flow into one per-agent CloudWatch log group by default.
That’s handy when you’re debugging agent behavior, since you don’t have to jump between multiple telemetry sources to connect the dots. It also helps scope IAM and CMK encryption to individual agents, which can reduce operational friction.
CloudWatch Logs now handles Application Load Balancer logs as vended logs
This should make real-time analysis, metric filters, and alerts easier to set up. In practice, that means less time spent troubleshooting traffic issues and less overhead managing separate log workflows.
AWS Glue Data Quality added distribution profiling
This makes it easier to spot skew and outliers early in the pipeline. Catching those issues sooner can help avoid expensive reprocessing later on.
AWS Glue Data Quality now supports anomaly detection with catalog result storage
That means rule outcomes and anomaly metadata are queryable instead of buried in a separate workflow. For teams managing data quality at scale, this can speed up regression detection and reduce downstream processing costs.
