Q1) What metrics does Lumen measure, and which are the primary decision makers?
Lumen analyzes four key performance metrics for intelligent tiering recommendations.
Primary decision makers:
IOPS (Input/Output Operations Per Second) – the main performance indicator determining workload requirements
Throughput – measures data transfer capacity and bandwidth needs
Secondary metrics:
Latency – used to assess performance impact after tier conversions
Cost – the primary optimization driver for all recommendations
This multi-dimensional analysis ensures that tier recommendations balance performance requirements with cost efficiency. IOPS and throughput patterns drive the technical feasibility of tier changes, while cost optimization remains the central objective of all recommendations.
What is the frequency of data collection and measurement?
Lumen operates on a daily collection cycle with comprehensive historical analysis capabilities:
Collection Frequency:
Daily snapshots: Metrics are collected once every 24 hours
Historical analysis windows: 30, 60, and 90-day lookback periods
Data availability dependency: Collection is subject to cloud provider metric availability
Analysis Methodology:
Lumen evaluates usage patterns across these extended timeframes to identify consistent trends
The platform distinguishes between temporary spikes and sustained usage patterns
Recommendations are based on statistically significant data points rather than isolated incidents
This approach ensures that tier recommendations are based on genuine workload patterns rather than temporary fluctuations, providing enterprise-grade reliability for critical storage decisions.
Q2) How does Lumen decide on recommendations?
Lumen's recommendation engine employs a sophisticated decision-making process that prioritizes cost optimization while ensuring performance requirements are met:
Primary Decision Framework:
IOPS Analysis: Evaluates whether current tier IOPS capacity matches actual usage patterns
Throughput Assessment: Determines if bandwidth requirements align with tier specifications
Cost Optimization: Identifies opportunities to reduce spending while maintaining workload performance
Recommendation Types:
Downtiering: Recommended when disks are over-provisioned, leading to unnecessary spend
Uptiering: Suggested when a tier above is cheaper and provides better performance at lower cost.
Cost-Driven Philosophy:
Every recommendation is fundamentally driven by cost optimization. Lumen ensures that enterprises pay only for the performance they actually utilize, eliminating the common practice of defaulting to premium tiers due to uncertainty about requirements.
Q3) How does Lumen account for workload cyclicity?
Lumen's data collection methodology is specifically designed to capture and analyze cyclical workload patterns:
Cyclicity Detection:
Extended observation periods: 30, 60, and 90-day analysis windows capture seasonal and cyclical patterns
Pattern recognition: Identifies recurring usage spikes, seasonal variations, and business cycle impacts
Trend analysis: Distinguishes between cyclical patterns and one-time events
Adaptive Recommendations:
Recommendations account for peak usage periods within cyclical patterns
The platform ensures tier selections accommodate regular high-demand cycles
Historical data analysis prevents under-provisioning during predictable peak periods
This approach ensures that tier recommendations accommodate your organization's specific operational cycles, preventing performance issues during anticipated high-usage periods.
Q4) How does Lumen handle IOPS and throughput bursts?
Lumen employs advanced statistical analysis to handle burst scenarios and prevent under-provisioning:
Burst Analysis Methodology:
90th percentile analysis: Recommendations are based on the 90th percentile of usage data points
Statistical significance: Ensures that occasional spikes don't drive unnecessary over-provisioning
Data point volume: Comprehensive analysis across thousands of daily data points
Burst Accommodation:
Tier recommendations account for legitimate burst requirements while filtering out anomalous spikes
The 90th percentile approach ensures that 90% of your workload demands are comfortably met
Prevents both under-provisioning (which causes throttling) and excessive over-provisioning (which wastes money)
This statistical approach provides enterprise-grade reliability while optimizing costs, ensuring your applications perform optimally during both normal operations and legitimate burst scenarios.
Q5) What is the cost impact of Azure API calls, and can they be limited?
The cost impact of Azure Monitor API calls for Lumen's operations is negligible and designed to minimize billing impact:
Cost Structure:
Azure Monitor pricing: First 1,000,000 API calls per month are free
Additional calls: Minimal cost per 1,000 API calls
Lumen's daily collection frequency minimizes API call volume
Cost Optimization Features:
Efficient data collection: Daily snapshots rather than continuous polling
Batch operations: Optimized API usage patterns
Targeted metrics: Focus on essential performance indicators only
Practical Impact:
For typical enterprise deployments, Azure API costs represent less than 0.1% of overall storage costs. The massive savings from optimal tier selection (up to 70% cost reduction) far exceed any monitoring costs.
Reference: Azure Monitor Pricing Details
Q6) How does Lumen achieve zero-downtime tier conversions?
Lumen's breakthrough zero-downtime capability leverages Lucidity's proprietary AutoScaler technology:
Technical Implementation:
Seamless tier switching: Utilizes advanced storage virtualization techniques
No VM reboots required: Maintains continuous application availability
Live migration: Data remains accessible throughout the conversion process
Q7) What are the use cases for disk tier upgrades vs. downgrades?
Lumen's recommendations are cost-driven but address both performance optimization and waste elimination:
Upgrade Scenarios (5% of disks):
Performance bottlenecks: When current tiers cause application throttling
IOPS limitations: Insufficient I/O capacity affecting user experience
Throughput constraints: Bandwidth limitations impacting data processing
Business growth: Increased workload demands requiring higher performance
Higher tier at lower cost:Given dynamic pricing and features a higher tier may full-fill workload need at lower cost
Downgrade Scenarios (45% of disks):
Over-provisioning: Premium tiers selected "just in case" but not utilized
Changed workloads: Applications with reduced performance requirements
Cost optimization: Significant savings opportunities with minimal performance impact
Workload migration: Legacy tier selections no longer appropriate
Primary Driver: Cost Optimization
While performance is crucial, every Lumen recommendation fundamentally aims to optimize costs. The platform ensures you pay only for the performance you actually need, whether that means upgrading to prevent costly downtime or downgrading to eliminate unnecessary spend.
Q8) Why isn’t Lumen a one-time optimization?
Lumen provides continuous optimization rather than one-time fixes because cloud environments are inherently dynamic:
Dynamic Environment Factors:
Workload evolution: Application requirements change over time
Business growth: Data volumes and performance needs scale
Seasonal variations: Usage patterns fluctuate with business cycles
Technology updates: New cloud services and tier options become available
Continuous Monitoring Benefits:
Ongoing cost optimization: Identifies new savings opportunities as they emerge
Performance trend analysis: Detects gradual changes in application behavior
Proactive recommendations: Identifies cost leakage before it becomes a problem.
Real-Time vs. Historical Balance:
While more frequent monitoring could provide real-time insights, the current daily collection frequency balances optimization effectiveness with monitoring costs. This approach captures meaningful trends while maintaining cost efficiency.
Q9) What is the typical time required for disk tier conversions?
Conversion timing varies based on several factors, with dashboard visibility throughout the process:
Conversion Duration Factors:
Disk size: Larger volumes may require additional processing time
Cloud provider: Azure, AWS, and GCP have different underlying mechanisms
Tier type: Specific source and destination tier combinations
Current workload: Active I/O during conversion may extend timing
Zero Downtime Guarantee:
Regardless of conversion duration, applications remain fully operational throughout the entire process, eliminating the business impact traditionally associated with storage tier changes.
Q10) What is the performance impact during tier conversions?
Lumen guarantees zero performance degradation during tier conversion operations:
Performance Assurance:
No latency impact: Applications maintain consistent response times
Uninterrupted I/O: Continuous read/write operations throughout conversion
Transparent process: Applications are unaware that tier conversion is occurring
Technical Safeguards:
Live data migration: Advanced algorithms ensure seamless data movement
Rollback capability: Immediate restoration if any issues are detected
Monitoring and Validation:
Automated alerts: Immediate notification if any anomalies are detected in conversion
Success validation: Comprehensive verification of conversion completion
Enterprise-Grade Reliability:
Lucidity's zero-downtime technology has been tested across Fortune 500 environments, demonstrating consistent performance maintenance during tier conversions. This breakthrough eliminates the traditional trade-off between cost optimization and operational risk.
Q11) How did Lumen help Dometic overcome cloud storage inefficiencies?
Dometic faced severe cloud storage inefficiencies, with 97% of disks on incorrect tiers, wasted spend, slow manual tiering requiring downtime, and poor visibility. Lumen provided data-driven, no-downtime, one-click tiering recommendations, enabling $14K in additional annual savings from 100 disks, boosting tiering efficiency from 44% to 97%, and allowing cloud ops teams to make informed storage decisions in seconds without any performance impact
Q12) What are the future features in the roadmap for Lumen?
Lucidity is expanding Lumen into a full cloud storage intelligence platform with three key upcoming capabilities:
Waste Disk Identification
Automatically detects orphaned volumes, stopped-instance disks, unmounted storage, and attached-but-unused volumes.
Helps eliminate 40–60% of wasted capacity by pinpointing truly idle resources.
Non-Autoscaler Environment Recommendations
Recognizing that not all enterprises can immediately implement Lucidity's AutoScaler technology, the roadmap includes enhanced recommendation capabilities for traditional cloud environments through one time assessment.
Advanced Performance Monitoring
Offers predictive analytics to forecast capacity needs and avoid bottlenecks.
Correlates application behavior with storage metrics for impact-focused optimization.
Delivers multi-horizon trend insights to address seasonal patterns and growth.
Q13) Can you give an example where upgrading helps reduce cost?
This cost-saving opportunity occurs when high transaction volumes make a lower-tier disk more expensive than upgrading to a higher tier with no transaction fees. Lucidity has identified this pattern across multiple enterprise customers, delivering cost savings through strategic tier upgrades.
Real-World Example:
A customer running analytics workloads on 64 GiB Standard SSD E6 disks was generating 25 million transactions monthly per disk:
Before optimization (Standard SSD E6):
Base cost: $4.80/month
Transaction fees: (25,000,000 ÷ 10,000) × $0.002 = $5.00/month
Total cost: $9.80/month
After Lumen optimization (Premium SSD v2):
Flat rate: 64 GiB × $0.081 = $5.18/month
Monthly savings per disk: $4.62 (47% reduction)
Across 100 of such disks, this delivers $46,200 annual savings
Q14) We are managing infrastructure via IaaC tool like Bicep. If Lumen makes disk tier change outside Bicep how do we manage the drift occurring in Bicep
Lucidity manages the lifecycle of the disk seamlessly by adding lifecycle hooks to IaaC tools like terraform, bicep etc. This ensure that any change made by Lucidity or Lumen is not introducing a drift in the state of the IaaC tool
Q15) Can we revert the changes we have made through Lumen Disk Tiering at a later date?
Yes, changes can be reverted; however, this currently requires submitting a Customer Support (CS) request. Our engineering team will perform the reversal from the backend. This functionality is not yet available directly through the Lumen dashboard but is under development. The ETA for the feature is TBD.
Q16) The disks listed as unmounted on Lumen’s idle disks module are showing up perfectly fine and attached on Azure, why?
Those disks are indeed attached to VMs. However, our system flags them as “Unmounted disks”, which is slightly different from “Unattached disks.”
What “Unmounted” means
A disk is considered unmounted when it is physically attached to a VM, but the operating system on the VM is not using the disk. In this state, the disk:
1. has no mount point or drive letter assigned,
2. is not part of the filesystem, and
3. shows no read/write activity.
How this differs from “Unattached”
Unattached disks are not attached to any VM. These disks can be tracked on the "Unattached" section on the Idle Disks module on Lumen.
Unmounted disks are attached to a VM, but are not mounted on the VM’s filesystem. These disks can be tracked on the "Attached but unused" section on the Idle Disks module on Lumen.
Why this matters
Despite being attached to the VM, these disks are unused, and incur storage costs, so they’re often good candidates for review or cleanup.
Q17) What is Pricing for Lucidity Lumen? Is it included with Lucidity AutoScaler?
Lucidity Lumen is a separate product from Lucidity AutoScaler and can be considered an add-on. It follows a flat pricing model. Please contact your account executive for detailed pricing information.
(Backup) Where to see the metrics in Azure portal
Navigate to the VM being monitored and select Monitor -> Metrics
From the metrics, choose “Data Disk IOPS Consumed Percentage”
For aggregation choose “Max”
Click on “Apply Splitting”
From the menu for values choose “LUN”
Click on “Add Filter” at the top
From the list of LUNs choose only the LUNs to be monitored
Choose the time-frame over which metrics have to be monitored
For observing the throughput metrics, follow steps 1 through 8 and in step 2, instead of “Data Disk IOPS Consumed Percentage” choose “Data Disk Bandwidth Consumed Percentage”