AWS

Amazon RDS and Aurora have extended support for R8gd and M8gd instance families to 12 and 6 additional regions respectively. These instances leverage Optimized Reads, which utilizes local NVMe SSDs for ephemeral data like temporary tables to reduce network I/O. This architecture delivers up to 165% higher throughput and 120% better price-performance compared to R6g instances for Aurora PostgreSQL.

• R8gd and M8gd instances are now available in 12 and 6 new regions for PostgreSQL, MySQL, and MariaDB.
• Optimized Reads uses local NVMe storage for temp tables, cutting network latency for complex queries.
• Expect up to 165% throughput gains and 120% price-performance improvements over R6g instances.
• Plan migrations to these new regions to leverage lower latency and higher throughput for workloads.

Fleet impact: For Aurora PostgreSQL and RDS for PostgreSQL/MySQL/MariaDB fleets, prioritize migrating high-throughput or complex query workloads to R8gd/M8gd in the newly supported regions to capitalize on the 165% throughput gain. Verify that your application handles the local NVMe ephemeral storage constraints for temp tables correctly during failovers or instance replacements.

AWS has made Graviton4-based R8g instances generally available for Aurora and RDS across multiple new global regions, including parts of Asia, Europe, and GovCloud. Concurrently, M8g instances are now supported for standard RDS engines in additional regions such as Canada West, Africa, and South America. These updates extend the availability of next-generation ARM-based hardware designed for improved performance and cost efficiency.

• R8g instances are now GA for Aurora and RDS in Hyderabad, Melbourne, Zurich, London, Paris, GovCloud, Sao Paulo, and Mexico Central.
• M8g instances are supported for RDS in US West, Mumbai, Sydney, Seoul, Calgary, Milan, Cape Town, and others.
• Graviton4 instances deliver up to 40% performance gains and 29% better price/performance over Graviton3 for on-demand workloads.
• Engine support for R8g includes Aurora MySQL/PostgreSQL and RDS PostgreSQL, MySQL, and MariaDB.

Amazon RDS now permits up to four storage volume changes per instance within a rolling 24-hour window, eliminating the previous six-hour cool-off period. This update applies to supported engines including PostgreSQL, MariaDB, and others, allowing immediate follow-up adjustments once the prior optimization completes. The change enables faster scaling of capacity, type, or performance during sudden workload spikes without requiring downtime.

• Remove the 6-hour wait between storage changes; start new mods immediately after completion.
• Scale IOPS, throughput, or volume size up to four times daily for rapid response.
• Applies to RDS PostgreSQL, MariaDB, and other supported engines automatically.
• Maintains zero-downtime operations while adjusting storage performance parameters.

Amazon Redshift has made two new Graviton-powered instance sizes, rg.large and rg.12xlarge, generally available. These nodes offer up to 2.4x faster query performance compared to previous-generation RA3 instances while costing 30% less per vCPU. This expansion provides greater flexibility for right-sizing provisioned clusters to match specific workload requirements.

• New rg.large and rg.12xlarge sizes expand right-sizing options for provisioned Redshift clusters.
• Graviton-based performance is up to 2.4x faster than RA3 at 30% lower price per vCPU.
• Availability is restricted to the current P202 track; P201 users retain rg.xlarge/4xlarge.
• Migration from existing RA3 clusters is supported via Snapshot/Restore, Elastic, or Classic Resize.