AWS FinOps Updates 2026-05-07

CloudWatch gets easier to set up, so your monitoring stays consistent

Amazon CloudWatch just added a visual way to configure the CloudWatch agent in the EC2 console. Instead of editing JSON by hand, you can now use a visual configuration editor to deploy metrics, logs, and tracing settings.

That’s a pretty practical win for teams trying to keep observability consistent across fleets. It reduces configuration drift and cuts down the operational overhead that comes with managing agent settings one instance at a time.

OpenSearch makes it easier to spot and trim unused storage

Amazon OpenSearch Service added a new Unused Index insight in Cluster Insights. It detects indices with zero activity over 30 days and points you toward warm or cold tiers.

The insight is available across Regions and works with both the console and EventBridge. That makes it easier to turn the recommendation into action without extra detective work.

RDS for SQL Server gets more scale and better restore options

Amazon RDS for SQL Server now supports read replicas and cross-account snapshot sharing for instances using additional storage volumes. The feature supports instances with up to 256 TiB of total storage.

This is a strong update for large databases where read offloading and disaster recovery planning both matter. Read replicas can take pressure off primary instances, while cross-account snapshot sharing helps with backup, restore, and diagnostics workflows.

For FinOps and DR planners, it’s useful because it supports scalable replication and restore patterns without changing storage layouts. That can make large estate management more practical and cost-aware.

New RDS instance types give SQL Server more price-performance options

Amazon RDS for SQL Server added support for M8i and R8i instance families. AWS says they offer up to 15% better price-performance and higher memory bandwidth than prior Intel generations.

This matters for teams trying to balance SQL Server performance with spend. More instance choices usually mean more room to right-size workloads instead of overpaying for capacity that isn’t needed.

The update also gives customers a path to migrate existing instances to these families. That makes it easier to tune cost and performance together instead of treating them as separate goals.

Cross-region OpenSearch access cuts duplication and keeps dashboards centralized

Amazon OpenSearch UI now supports cross-region data access. A single UI can query domains across multiple Regions and accounts.

This is handy for teams that want a centralized view without copying data around. It lets you keep data where it already lives while still building cross-region dashboards and analytics.

CloudWatch Logs Insights can now query by tag, not just by name

Amazon CloudWatch Logs Insights added support for querying log groups by tags. You can now run queries across dynamically grouped log sets like Environment=prod.

That’s especially useful when log groups change often or span multiple accounts and Regions. You don’t have to keep updating long lists of names just to run a query.

For FinOps, this helps with cost-targeted analysis and log hygiene work. Tag-based queries make it easier to find and inspect the right logs without extra management overhead.

IAM quotas got bigger, which should reduce scaling friction

AWS IAM increased several account quotas, including roles and customer managed policies per account. OIDC providers also increased.

This update is less flashy, but it matters for large organizations scaling identities and service principals. Higher defaults mean fewer quota-related interruptions when workloads grow.

The result is less manual quota-request work and less operational overhead. That helps teams keep identity management from becoming a bottleneck during expansion.

The AWS Console Mobile App makes alarm investigations faster

The AWS Console Mobile App now includes CloudWatch Alarms investigation tools. You get interactive graphs, AI log summaries, and natural-language log search.

This is a nice on-call upgrade for teams that need to triage incidents while away from a desk. It shortens the path from alert to root-cause analysis.

EKS can now use local instance storage more easily

The EC2 Instance Store CSI driver is now generally available as an Amazon EKS add-on. That makes it simpler to use ephemeral local NVMe storage from Kubernetes.

This is a good fit for transient, data-heavy workloads that can benefit from local performance. Instance store can be lower cost and faster than networked EBS in the right scenarios.

So if you’ve got workloads that don’t need persistent block storage, this gives you a cleaner way to place them on the appropriate instance types. That can improve both performance and cost efficiency.

CloudFront cache invalidation gets more precise

Amazon CloudFront now supports invalidation by cache tag. You can tag cached objects and invalidate groups by tag instead of listing URLs or using broad wildcards.

That makes cache management more practical for larger sites and apps. More precise invalidation means less accidental overreach when you only need to refresh part of the cache.

It can also reduce unnecessary edge invalidations and egress. In cost terms, that helps keep origin traffic and cache-miss-driven spend under control.

ElastiCache gets better visibility into memory and network behavior

Amazon ElastiCache added 13 new CloudWatch metrics. The metrics cover areas like network baseline, allocator fragmentation, and blocked or rejected connections.

This is useful when you need to understand whether a cache is being throttled or fragmenting memory. Instead of relying on extra probes, you now get more signal directly in CloudWatch.

The FinOps benefit is better capacity planning and more cost-aware scaling. Since the metrics are available at no extra cost in all commercial Regions, they’re easy to start using.