Core Concepts

This page explains the key concepts you will encounter when using CLARITY. Understanding these fundamentals will help you get the most out of the platform.

Providers

CLARITY supports three cloud providers with consistent feature coverage:

Provider	Billing Source	Resource Discovery	Commitment Types
AWS	Cost Explorer API	EC2, RDS, S3, ECS, EKS, Lambda, and more	Reserved Instances, Savings Plans
Azure	Cost Management API	VMs, SQL, Storage, AKS, App Services, and more	Reservations
GCP	BigQuery Billing Export	Compute, Cloud SQL, GKE, Cloud Storage, and more	Committed Use Discounts

You can connect multiple accounts per provider. Each account syncs independently, and all data is presented in a unified view with provider-level filtering.

Credentials

CLARITY connects to your cloud accounts using read-only credentials:

• AWS: IAM user access keys with Cost Explorer, EC2, and CloudWatch read permissions
• Azure: Service Principal with Reader and Cost Management Reader roles
• GCP: Service Account with Viewer and Billing Account Viewer roles

Security

All credentials are encrypted at rest using AES-256 encryption with per-credential key derivation. Credentials are never logged, never displayed in the UI after entry, and never transmitted to external services.

CLARITY never creates, modifies, or deletes any resources in your cloud accounts. All access is strictly read-only.

Sync

CLARITY automatically collects data from your cloud providers on a configurable schedule. Each sync cycle pulls:

• Billing data — Service-level and resource-level cost breakdowns
• Resource inventory — Complete hierarchy of active resources
• Performance metrics — CPU, memory, network, and storage utilization
• Provider recommendations — Native optimization suggestions from AWS, Azure, and GCP
• Commitment details — Active reservations, savings plans, and committed use discounts

TIP

The first sync runs immediately after you add credentials. Subsequent syncs run automatically at the configured interval. You can also trigger a manual sync at any time from the provider settings.

After each sync, CLARITY automatically reconciles stale resources — if a resource no longer appears in your cloud account, it is marked accordingly in the inventory.

Cost Allocation

CLARITY uses a 4-strategy allocation chain to assign costs to every resource in your inventory. Each strategy is tried in order, and the first one that produces a result is used:

Strategy 1: Direct Billing

Costs are pulled directly from your cloud provider's billing APIs (AWS Cost Explorer, Azure Cost Management, GCP BigQuery). This is the most accurate source and is used whenever available.

Highest accuracy

Strategy 2: Dynamic Pricing

When direct billing data is not available at the resource level, CLARITY queries live pricing APIs for your provider and region. Pricing data is cached and refreshed periodically, with static fallback tables as a safety net.

API-based pricing

Strategy 3: Proportional Allocation

For resources that are part of a service (e.g., individual containers within a cluster), CLARITY takes the known service-level cost and distributes it proportionally across child resources based on their relative size and utilization.

Service-level split

Strategy 4: Kubernetes Namespace Allocation

For Kubernetes workloads, costs are allocated at the namespace level based on actual CPU and memory consumption relative to total cluster capacity.

K8s native

Every cost figure in CLARITY includes a source indicator so you always know which strategy produced it.

Insights vs Recommendations

CLARITY distinguishes between two types of optimization signals:

	Insights	Recommendations
What	Observations about resource behavior	Specific actions to reduce cost
Example	"This VM averaged 3% CPU over 14 days"	"Downsize from m5.xlarge to m5.large to save $85/month"
Source	CLARITY's analysis engine	Cloud provider APIs + CLARITY rules
AI Validation	Verified by AI when enabled	Verified by AI when enabled

Both insights and recommendations support on-demand AI explanations — click the Explain button on any item for a detailed, context-aware analysis of why it was flagged and what the potential impact of acting on it would be.

Classification

Insights are categorized by severity:

• Critical — Immediate action recommended (e.g., idle resources costing >$50/month)
• Warning — Optimization opportunity worth investigating
• Info — Informational observation, no immediate action needed

Anomalies

CLARITY continuously monitors your cloud spending patterns and flags statistical anomalies — cost spikes or drops that deviate significantly from historical baselines.

When an anomaly is detected:

1. The deviation is quantified (percentage above/below expected spend)
2. The affected service and region are identified
3. If AI analysis is enabled, an automatic root cause explanation is generated
4. Configurable alerts can notify your team via email

Anomaly detection uses rolling historical baselines, so seasonal patterns and expected growth are accounted for — you only get alerted on genuinely unusual spend changes.

Forecast

CLARITY projects your future cloud spend using multiple forecasting methods:

• Provider-native forecasts — AWS and Azure offer built-in forecasting APIs
• Regression-based projections — Statistical models fit to your historical billing data
• Trend extrapolation — Short-term projections based on recent spend velocity

Forecasts are calculated per provider and aggregated into a unified view. They account for actual billing days (not calendar days) and exclude incomplete data from the current day to avoid skewing projections.

TIP

Forecasts become more accurate as CLARITY accumulates more historical data. At least 14 days of billing history is recommended for reliable projections.

Tiers

CLARITY uses a subscription tier system to control feature access and resource limits. Your tier determines:

Capability	What It Controls
Cloud accounts	Maximum number of connected provider accounts
Users	Maximum number of team members
Budgets	Number of budget alerts you can configure
Commitments	Access to RI/SP/CUD recommendation engine
Cost Allocation	Access to chargeback and showback features
AI Insights	Access to AI-powered analysis and explanations
Reports	Report generation and scheduled delivery

Your current tier is displayed in the sidebar. Feature-gated capabilities show a clear indicator when they require a higher tier, so you always know what is available and what requires an upgrade.

Next Steps

• Quickstart Guide — Connect your first cloud account
• Dashboard — Learn about the main dashboard
• Cloud Setup Guides — Detailed credential setup per provider