Overview

The Cloud Cost Recommendation section provides a comprehensive view of cloud cost optimization opportunities across AWS, Azure, and GCP. It automatically analyzes cloud resource usage to identify inefficiencies such as idle instances, over-provisioned storage, underutilized clusters, and obsolete resources. Each recommendation card highlights potential savings, affected resources, and suggested corrective actions to reduce unnecessary expenses.

Supported Technology

AWS: Enables monitoring and configuration of Amazon Web Services resources across compute, storage, networking, and cost visibility.
GCP: Provides visibility into Google Cloud Platform workloads, offering performance, usage, and configuration insights.
Azure: Supports tracking and managing Azure cloud services, including performance metrics and cost visibility setup.
On-Prem: Allows monitoring of on-premise infrastructure such as servers, VMs, networks, and storage systems.
Containers: Offers monitoring of containerized environments like Docker and Kubernetes, covering pods, clusters, and resource usage.

Users can select a particular technology to filter out the recommendations.

Cost Summary

Total Savings: Displays the overall cost savings identified through optimization recommendations.
Total Investments: Shows the total estimated cost required to implement the suggested optimizations.

View Recommendation Options

Search by Title or Content: Allows you to quickly find recommendations by typing any part of the recommendation name or related content.
Cards View: Displays recommendations in a card-based layout, showing key highlights like severity, affected items, and potential savings in a visually compact format.
Table View: Presents recommendations in a detailed tabular format, allowing sorting, filtering, column management, and exporting for deeper analysis.

Common Recommendation Options

Each recommendation card displays associated resources and an option to apply the recommendation. Users can access additional Settings and Exclude options to customize or ignore recommendations.

Menu Options

Settings: This option allows users to customize how recommendations are generated and displayed. From here, users can adjust evaluation parameters such as activity duration or utilization thresholds, select specific services to include or exclude from analysis, and choose an action mode like “Recommended.” These settings help tailor the cost optimization suggestions according to user preferences and ensure that future recommendations align with the selected criteria.
- Example: In the Over-Provisioned EBS IOPS recommendation, you can change the utilization threshold (e.g., 80%) and select the action mode as Recommended. The new settings apply from the next recommendation check.
Exclude Pools: Allows users to exclude specific resource groups or pools from recommendation analysis, preventing them from appearing in future cost optimization results.
Pins: Allows users to pin important recommendation cards so they remain visible at the top for easy access and quick review.

Recommendation

Over-Provisioned IOPS Highlights storage volumes whose provisioned IOPS exceed observed read/write usage, resulting in unnecessary cost. The card lists affected volumes and an estimated monthly saving for each. Suggested action: Right-size provisioned IOPS to the actual workload or migrate to a volume type that decouples IOPS from capacity. (example: rule might flag volumes using less than 80% of provisioned IOPS.)
Instances with Spot / Preemptible Opportunities: Identifies instances with intermittent or non-critical usage that are suitable candidates for Spot (AWS) or Preemptible (GCP/Azure) pricing. The card shows impacted instances and potential savings. Suggested action: Migrate tolerant workloads to Spot/Preemptible instances, use Auto Scaling Groups, or schedule non-critical tasks to exploit lower prices. (example: rule might flag instances that have run only a few hours per day over a recent week.)
RI / SP Opportunities: Detects compute usage patterns that are not sufficiently covered by Reserved Instances or Savings Plans. The card provides suggested reservation or savings plan adjustments and estimated savings. Suggested action: Evaluate adding or adjusting RI/Savings Plan coverage to reduce on-demand spending.
Instances for Shutdown: Identifies instances with sustained low CPU and network activity, indicating little or no usage. The card lists such idle instances along with potential savings. Suggested action: Create shutdown schedules or decommission unused instances. (example: rule may flag instances showing consistently low activity over several days.)
Obsolete Snapshot Chains: Detects snapshot chains that have not been used to create new volumes for a long duration. The card displays these chains and their potential savings. Suggested action: Delete unused snapshot chains to reclaim storage. (example: rule may flag chains inactive beyond retention policies.)
Obsolete Snapshots with No Volume: Highlights snapshots not associated with any active volume, resulting in avoidable storage cost. The card lists such snapshots with monthly savings. Suggested action: Remove snapshots no longer serving operational needs. (example: rule may flag snapshots unreferenced for extended periods.)
Abandoned Instances: Identifies running instances with extremely low CPU and network activity, indicating abandonment. The card displays these instances with estimated savings. Suggested action: Terminate or repurpose idle compute resources. (example: rule may flag instances showing negligible activity for prolonged durations.)
Underutilized Instances: Detects instances operating below optimal capacity, presenting rightsizing opportunities. The card lists such instances with potential savings. Suggested action: Resize to a smaller instance type while maintaining performance needs. (example: rule may flag instances with low peak CPU utilization.)
Unattached Volumes: Identifies storage volumes not attached to any instance but still incurring cost. The card displays all unattached volumes with savings potential. Suggested action: Delete or repurpose unused volumes. (example: rule may flag volumes detached for an extended time.)
Migrate Volumes: Highlights older-generation volume types that can be migrated to newer, cost-efficient classes. The card lists such volumes with potential savings. Suggested action: Migrate GP2 or equivalent to GP3 or modern alternatives. (example: rule may flag volumes running on older types.)
Over-Provisioned Throughput: Detects volumes where provisioned throughput significantly exceeds observed read/write throughput. The card displays affected resources and savings. Suggested action: Adjust throughput to match workload needs. (example: rule may flag volumes consistently underusing provisioned throughput.)
Unused Instance Running Redis: Identifies Redis instances with no keyspace activity and not part of any cluster role. The card lists unused Redis nodes along with savings. Suggested action: Shut down or remove inactive Redis deployments. (example: rule may flag nodes with zero cache activity.)
Unused Instance Running Memcached: Highlights Memcached instances with no cache operations and no active role. The card lists inactive instances and savings. Suggested action: Terminate or remove unused Memcached nodes. (example: rule may flag nodes with no key/value activity.)
Unused Instance Running Postgres: Detects PostgreSQL instances with almost no active connections or workload. The card lists such instances with savings potential. Suggested action: Stop or decommission unused databases. (example: rule may flag instances with negligible session activity.)
Unused Instance Running MySQL: Identifies MySQL instances with minimal or no activity. The card displays such instances with savings. Suggested action: Remove or downsize inactive MySQL deployments. (example: rule may flag servers with few or no active connections.)
Over-Provisioned Instance Running Redis: Highlights Redis instances over-allocated in compute relative to actual usage. The card lists these nodes with savings. Suggested action: Resize to a smaller instance class. (example: rule may flag nodes with consistently low CPU usage.)
Over-Provisioned Instance Running Memcached: Detects Memcached instances using far less compute than provisioned. The card lists such instances with potential savings. Suggested action: Downsize to a smaller instance class. (example: rule may flag nodes with low CPU consumption.)
Over-Provisioned Instance Running Postgres: Identifies PostgreSQL servers with low CPU and large available memory, indicating oversized resources. The card lists candidates for rightsizing. Suggested action: Move to a smaller instance class. (example: rule may flag databases with low resource utilization.)
Over-Provisioned Instance Running MySQL: Highlights MySQL instances with over-provisioned compute or memory. The card lists opportunities to reduce cost through downsizing. Suggested action: Right-size MySQL servers to match the workload. (example: rule may flag instances with excess unused resources.)
S3 Lifecycle Configuration: Detects buckets missing lifecycle rules to manage noncurrent or aged objects. The card lists such buckets with savings potential. Suggested action: Implement policies to transition or expire outdated object versions. (example: rule may flag buckets storing many noncurrent versions.)
S3 Intelligent Tiering Configuration: Identifies buckets where Intelligent-Tiering is not enabled despite variable access patterns. The card shows such buckets with savings potential. Suggested action: Enable Intelligent-Tiering for cost-efficient storage. (example: rule may flag buckets with infrequently accessed objects.)
Extended Support Instance: Highlights RDS database engines running on deprecated or outdated versions that may incur extended support charges. The card lists such instances with potential savings. Suggested action: Upgrade to a supported engine version to eliminate extended support costs. (example: rule may flag databases running on versions past vendor support timelines.)
Purchase Reserved Instance (RDS): Detects RDS instances running continuously under on-demand pricing. The card lists such instances with reservation-based savings potential. Suggested action: Purchase Reserved Instances for predictable, long-running database workloads. (example: rule may flag RDS nodes active for extended periods.)
Old Snapshots: Identifies database snapshots retained beyond the defined retention or operational need. The card lists such snapshots and their savings. Suggested action: Delete outdated or redundant snapshots to reduce storage cost. (example: rule may flag snapshots older than policy limits.)
Underutilized Cluster: Highlights database or cache clusters with consistently low utilization. The card lists these clusters with potential savings.
Unused Cluster: Detects clusters with no recent client connections or workload activity. The card lists unused clusters with savings opportunities. Suggested action: Delete or decommission clusters not serving any operational need. (example: rule may flag clusters with zero active sessions.)
Purchase Redshift Reserved Instance: Identifies Redshift clusters running for long periods without RI commitments. The card shows reservation-saving opportunities. Suggested action: Purchase Reserved Instances for consistently running Redshift workloads. (example: rule may flag clusters active for months.)
Underutilized ElastiCache Cluster: Detects cache clusters with minimal compute or memory usage. The card lists such clusters with savings potential. Suggested action: Resize or retire underutilized ElastiCache clusters. (example: rule may flag clusters using only a small portion of their resources.)
Unused ElastiCache Redis Clusters: Highlights Redis clusters with no requests, replication, or data activity. The card lists unused clusters with cost savings. Suggested action: Remove or repurpose Redis clusters that show no activity. (example: rule may flag clusters with zero cache hits.)
Purchase ElastiCache Reserved Instance: Detects ElastiCache clusters running continuously under on-demand cost. The card lists nodes eligible for reservation savings. Suggested action: Purchase Reserved Instances to reduce long-running cache compute expense. (example: rule may flag cache nodes operating 24×7.)
Unused MQ Broker: Identifies message brokers with no active connections or traffic. The card lists idle brokers and potential savings. Suggested action: Stop or delete MQ brokers that are unused. (example: rule may flag brokers with zero client connections.)
Old ECR Images: Highlights container images in Amazon ECR that have not been used recently. The card lists such images with savings potential. Suggested action: Delete outdated or unused container images to reclaim storage. (example: rule may flag images unused for months.)
OpenSearch Cluster: Identifies OpenSearch clusters receiving no search or indexing activity. The card lists such clusters with cost savings. Suggested action: Decommission or downsize OpenSearch clusters that have no active usage. (example: rule may flag clusters with zero queries.)
Purchase OpenSearch RI: Detects OpenSearch instances running long-term with on-demand billing. The card lists RI savings opportunities. Suggested action: Purchase Reserved Instances for continuously running OpenSearch workloads. (example: rule may flag instances active for long stretches.)
ELB Low Request Count: Identifies Elastic Load Balancers receiving minimal incoming requests. The card lists such ELBs with savings potential. Suggested action: Remove or consolidate low-traffic load balancers. (example: rule may flag ELBs with only a few requests per day.)
Unused Classic Load Balancers: Highlights Classic Load Balancers without active backend instances or traffic. The card lists unused CLBs and savings. Suggested action: Delete CLBs that are no longer serving traffic. (example: rule may flag CLBs with no registered targets.)
Unused Network Elastic Load Balancer: Detects Network Load Balancers that are not processing traffic and still incurring charges. The card lists such NLBs and their savings. Suggested action: Remove or decommission unused NLBs. (example: rule may flag NLBs with zero flow data.)
Unused Application Load Balancer: Identifies Application Load Balancers that are not receiving traffic. The card lists idle ALBs and potential savings. Suggested action: Delete unused ALBs to avoid unnecessary cost. (example: rule may flag ALBs with zero requests.)
Unused NAT Gateway: Detects NAT Gateways with no network flow or data activity. The card lists such gateways and savings. Suggested action: Remove unused NAT Gateways when not required. (example: rule may flag NAT Gateways processing no bytes.)
Unused DynamoDB GSI: Highlights Global Secondary Indexes without any read/write activity. The card lists these GSIs with potential savings. Suggested action: Delete GSIs that applications are not using. (example: rule may flag indexes with zero requests.)
Delete Extra DynamoDB Backups: Identifies DynamoDB tables with more on-demand backups than required. The card lists such tables with savings. Suggested action: Delete unnecessary backups beyond defined retention needs. (example: rule may flag tables exceeding acceptable backup count.)
Over-Provisioned DynamoDB Capacity: Identifies DynamoDB tables where provisioned read/write capacity is significantly higher than actual consumption. The card lists these tables with potential savings. Suggested action: Right-size DynamoDB provisioned capacity based on real usage patterns. (example: rule may flag tables consuming only a fraction of their allocated capacity.)
Migrate DynamoDB to Standard Class: Highlights tables using specialized or higher-cost storage classes unnecessarily. The card lists such tables with savings potential. Suggested action: Migrate tables to the Standard class when advanced features are not required. (example: rule may flag tables with general-purpose access patterns.)
Migrate DynamoDB to Infrequent Access Class: Identifies tables with low read/write frequency that fit the Infrequent Access (IA) class. The card lists such tables with potential savings. Suggested action: Move low-access tables to the IA class to optimize storage and throughput cost. (example: rule may flag tables with periodic, low-volume requests.)
Obsolete IPs: Detects Elastic IPs that are unassociated with any instance or resource. The card lists these IPs with savings potential. Suggested action: Release unused Elastic IPs to avoid unnecessary charges. (example: rule may flag IPs unassigned for long durations.)
Not Attached Volumes: Highlights EBS volumes detached from EC2 instances but still incurring cost. The card lists such volumes with savings. Suggested action: Delete or repurpose unattached volumes. (example: rule may flag volumes detached beyond the allowed timeframe.)
Obsolete Images: Identifies Amazon Machine Images (AMIs) not used recently for instance launches. The card lists these unused images with savings potential. Suggested action: Delete obsolete AMIs along with related snapshots to free up storage. (example: rule may flag AMIs unused for several weeks.)
Reserved Instances Opportunities: Detects long-running EC2 instances operating without RI or Savings Plan coverage. The card lists such instances with cost-saving potential. Suggested action: Purchase Reserved Instances or Savings Plans for steady, predictable workloads. (example: rule may flag instances active for more than a typical RI threshold.)
Obsolete Snapshots: Highlights snapshots that are no longer required and continue to incur storage cost. The card lists these snapshots with estimated savings. Suggested action: Remove obsolete snapshots after validating backup requirements. (example: rule may flag snapshots unused beyond standard retention.)
Not Deallocated Instances: Detects stopped instances that still incur storage or IP-related charges because they are not fully deallocated. The card lists such instances with savings. Suggested action: Fully deallocate or delete stopped instances to avoid unnecessary cost. (example: rule may flag instances stopped for extended periods.)
Old Generation Instances: Identifies older-generation compute instances that offer lower performance and higher cost than current-generation types. The card lists these instances with savings opportunities. Suggested action: Migrate workloads to the latest-generation instance families for better cost-to-performance efficiency. (example: rule may flag instances running on legacy families.)
Instances with Subscription Opportunities: Highlights compute instances that can benefit from sustained-use or vendor subscription pricing. The card lists such instances with savings. Suggested action: Purchase applicable subscriptions for workloads with stable usage patterns. (example: rule may flag instances with consistently high monthly usage.)
Abandoned Kinesis Streams: Detects Kinesis data streams that are provisioned but not processing any data. The card lists these streams with savings potential. Suggested action: Delete or consolidate streams that have no active data ingestion. (example: rule may flag streams with zero operations over time.)
Abandoned Images: Identifies images unused for instance or volume creation. The card lists such images with related savings. Suggested action: Delete abandoned images no longer needed for deployments. (example: rule may flag images unused for several weeks.)
Inactive IAM Users: Highlights IAM accounts with no recent API or console activity. The card lists these users along with cleanup recommendations. Suggested action: Disable or remove inactive IAM users to enhance security and reduce account clutter. (example: rule may flag users inactive for long durations.)
IAM Users with Unused Console Access: Detects users who have not used their login password for an extended period. The card lists such users and potential actions. Suggested action: Disable or revoke unused console passwords to maintain security. (example: rule may flag accounts without console logins across multiple cycles.)
IO1 Volumes: Identifies high-performance IO1/IO3 volumes that can be migrated to GP3 for better cost efficiency without impacting required performance. The card lists such volumes with savings. Suggested action: Migrate IO1/IO3 volumes to GP3 unless the workload specifically requires provisioned IOPS. (example: rule may flag IO1 volumes underutilizing IOPS.)
Over-Provisioned EBS Volume IOPS: Detects EBS volumes with provisioned IOPS significantly higher than used. The card lists such volumes with savings potential. Suggested action: Right-size EBS IOPS to align with actual I/O demand. (example: rule may flag volumes consistently using far less IOPS.)
Over-Provisioned Storage: Highlights storage volumes with very low utilization compared to their allocated size. The card lists such volumes with potential cost savings. Suggested action: Resize, compress, or migrate volumes to smaller capacity tiers. (example: rule may flag volumes using only a small fraction of provisioned size.)