🔍

Scanning Tiers

Three Tiers of Visibility

EchelonGraph provides three scanning tiers — each adding deeper visibility into your infrastructure. Start agentless with Tier 1 and expand as your security posture matures.


Tier 1: EcheSky (Agentless Cloud Scanning)

Deployment: Agentless — API-only, no software to install Scan mode: On-demand or scheduled Best for: Immediate cloud visibility without touching infrastructure

What Tier 1 Discovers

CategoryWhat's Scanned
ComputeVMs, serverless functions, container services
NetworkVPCs, subnets, security groups, firewalls, load balancers
StorageObject stores and block storage — encryption and access policies
IdentityIAM users, roles, policies, service accounts
DataManaged databases — encryption, public access, backup status
CertificatesTLS certificates and expiry dates

What Tier 1 Detects

  • 440+ misconfiguration rules mapped to CIS v2.0 benchmarks
  • CVE correlation — automatic matching of service versions to known vulnerabilities
  • Attack path analysis — maps reachable paths from internet-facing nodes to sensitive data
  • Compliance scoring across all 9 supported frameworks

> ⚠️ Kubernetes visibility (capability boundary): Your Kubernetes cluster surfaces as a single asset — the cloud control-plane row (e.g., EKS_cluster:my-prod, GKE_cluster:prod-1, AKS_cluster:eu-west). Tier 1 sees what's *registered* with your cloud provider's API, not what's *running inside*. Namespaces, pods, services, RBAC objects, NetworkPolicies, and CRDs are invisible from outside the cluster — that's a fundamental capability boundary, not a licensing decision. For inside-cluster topology, deploy Tier 3.

How It Works

  1. Scanner authenticates with your cloud provider using read-only credentials
  2. Discovers all assets using API calls — no agents or disk access
  3. Evaluates 440+ misconfiguration rules and correlates known CVEs
  4. Builds attack path graph from internet-facing nodes inward
  5. Findings appear in your dashboard and trigger alerts automatically

Tier 2: EcheNet (Deep Network Scanning)

Deployment: Standalone container (Docker or Helm chart), Cloud Run Job, or Kubernetes Deployment/CronJob Scan mode: One-shot (deploy-and-destroy) or Daemon (persistent with health/metrics endpoint) Best for: Network-level visibility, container security, Kubernetes hardening, private subnet scanning

Capabilities Beyond Tier 1

FeatureDescription
Deep network scanningTCP port scanning with service detection (30+ protocols), SSL/TLS certificate analysis, HTTP fingerprinting, and DNS enumeration across private subnets
Shadow IT detectionIdentify unauthorized services — open databases (Redis, MongoDB, Elasticsearch without auth), exposed admin panels (Jenkins, phpMyAdmin, Kibana), and rogue APIs
Container image scanningAnalyze running containers and registry images for CVEs using integrated Trivy CLI
SBOM generationSoftware Bill of Materials in CycloneDX and SPDX formats for every container
Kubernetes securityCIS benchmark rules (12 rules), RBAC audit (5 rules), network policy analysis (3 rules), and K8s attack path graph generation
Runtime-enriched attack pathsAttack paths based on actual runtime state — scanned ports, exposed services, and live container topology
Config drift detectionDetect infrastructure drift between scan cycles — new hosts, changed ports, removed services
MITRE ATT&CK mapping14 technique mappings (T1046, T1190, T1530, etc.) applied to every finding
PCI DSS v4.0 mappingAutomated control mapping for payment card compliance
Agent enrollmentTwo-step cryptographic handshake — one-time enrollment token exchanged for a long-lived agent credential

Enterprise Features

FeatureDescription
BYOK encryptionAES-256-GCM encryption of all telemetry before transmission — SaaS backend never sees plaintext
Air-gapped modeZero outbound network — scan results written as SHA-256 integrity-verified JSON files to local disk
License validationHMAC-SHA256 signed license keys with feature gating, asset caps, expiry warnings, and Prometheus health metrics
Zero-downtime key rotationDual-key decryption window — rotate encryption keys without scan interruption

Tier 2 shows you what's actually running — not just what's configured.

> Kubernetes scope (Tier 2): CIS benchmark misconfig audit (12 rules), RBAC misconfig rules (5 rules), NetworkPolicy analysis (3 rules), and container image CVE scan via Trivy. Tier 2 reads via the K8s API for one-shot posture detection; it does not maintain a continuous in-cluster topology watch (no live event stream on namespaces / pods / RBAC objects / CRDs). Continuous topology + real-time RBAC inventory + shadow-secret detection + CRD-aware AI workload discovery require Tier 3.


Tier 3: EcheDeep (Runtime Telemetry)

Deployment: Kernel-level agent (Kubernetes DaemonSet) + SDK Scan mode: Continuous + event-driven Best for: Zero-day detection, lateral movement analysis, deep runtime security

Capabilities Beyond Tier 2

FeatureDescription
Live K8s topology — 16+ kindsclient-go informer watch on Namespace / Node / Pod / Service / Deployment / StatefulSet / DaemonSet / Job / CronJob / Ingress / NetworkPolicy / ServiceAccount / Role / ClusterRole / RoleBinding / ClusterRoleBinding / HPA / PVC / PV — plus CRD registry. Every change observed in real time (sub-second). Cloud flavour (GKE / EKS / AKS / on-prem) auto-detected from Node ProviderID so assets land under the correct cloud account in your inventory.
Strict-ZK Secret + ConfigMap inventoryUses Kubernetes metadatainformer so the API server returns only PartialObjectMetadata envelopes — the agent process never holds .Data bytes in memory by construction. Detect shadow secrets created outside your Vault / KMS provisioning flow without ever reading their values. The strictest secret-discovery posture in the industry.
Shadow AI / ML workload discoveryFirst-class CRD watch for Kubeflow Notebook, KServe InferenceService, Argo Rollouts, KubeRay RayCluster, Seldon SeldonDeployment, Run:ai RunaiJob. Surfaces the data-science team's GPU pod that bypassed your CI gate — the security blind spot every other CSPM misses.
End-to-end Zero-Knowledge with browser-side decryptSensitive findings (T3.4 process, T3.5 anomaly, T3.6 IOC) encrypted on-host with a per-event AES-256-GCM DEK wrapped under your KMS (AWS / GCP / Vault). EchelonGraph SaaS stores ciphertext only; the analyst's browser decrypts via our open-source TypeScript Browser SDK. The only commercial agent where the SaaS operator can't read your raw runtime data — Sysdig, Aqua, Wiz all decrypt server-side.
Shadow API discoveryDetect undocumented APIs communicating across your network
Zero-day correlationThreat intelligence integration via STIX 2.1 + TAXII 2.1 (abuse.ch URLhaus, Feodo Tracker, CISA KEV)
Lateral movement simulationModel how an attacker could move through your infrastructure
Kernel-level telemetryDeep process and system call monitoring via eBPF — XDP, TC, tracepoints
Predictive threat detectionML-based anomaly detection on runtime patterns (24h baseline + EWMA + seasonality)
Auto-remediation9 IaC patch templates (K8s NetworkPolicy / PodSecurity / RBAC / Capabilities + AWS Terraform SG / S3 / IAM / CloudTrail / RDS) with GitHub PR + Slack approval flow

Tier Comparison Matrix

CapabilityTier 1Tier 2Tier 3
Cloud resource discoveryYesYesYes
Network topology mappingYesYesYes
IAM and identity auditYesYesYes
440+ misconfiguration rulesYesYesYes
CVE correlationYesYesYes
CIS benchmark mappingYesYesYes
3D blast radius visualizationYesYesYes
Compliance scoring (9 frameworks)YesYesYes
Deep network scanning (TCP/SSL/HTTP/DNS)YesYes
Shadow IT detectionYesYes
Container image scanningYesYes
SBOM generationYesYes
Kubernetes CIS + RBAC auditYesYes
Runtime-enriched attack pathsYesYes
Config drift detectionYesYes
MITRE ATT&CK mappingYesYes
PCI DSS v4.0 control mappingYesYes
BYOK encryptionYesYes
Air-gapped modeYesYes
K8s cluster as a single asset (cloud control plane)YesYesYes
K8s namespaces / pods / services as individual assetsYes
K8s RBAC inventory (Roles, ClusterRoles, bindings, ServiceAccounts) — liveYes
K8s Secret + ConfigMap inventory (metadata-only, zero-data-leak)Yes
Shadow AI/ML workload discovery (Kubeflow, KServe, Argo Rollouts, KubeRay, Seldon CRDs)Yes
Real-time eBPF runtime detection (process, anomaly, threat-intel)Yes
End-to-end Zero-Knowledge with browser-side decryptYes
Shadow API discoveryYes
Zero-day threat intelligenceYes
Lateral movement simulationYes
Kernel-level telemetryYes

Which Tier Is Right for You?

Use CaseRecommended
On a budget — need cloud security without spending todayTier 1 (Free Forever — 3 cloud accounts, 500 assets, 90-day retention)
Quick cloud posture assessmentTier 1
Compliance audit preparationTier 1
Private subnet and network scanningTier 2
Container and Kubernetes posture auditTier 2
Runtime vulnerability detectionTier 2
Regulated industry (finance, healthcare, government)Tier 2 or 3
Air-gapped / self-hosted deploymentTier 2
Live Kubernetes inventory — namespaces, pods, RBAC, NetworkPoliciesTier 3
Shadow AI / ML workload discovery (Kubeflow, KServe, Ray, Argo)Tier 3
Shadow secret detection without reading secret valuesTier 3
Zero-Knowledge encryption with browser-side decryptTier 3
Zero-day threat detectionTier 3
Advanced threat huntingTier 3

All tiers include the full compliance engine, blast radius visualization, and alerting capabilities. Higher tiers add deeper data sources — they build on lower tiers rather than replacing them. Tier 1 is free forever (3 cloud accounts, 500 assets, 90-day retention) — start there if you're shopping for the best cloud security platform on a budget; upgrade to Tier 2 ($49/node/mo) for network depth or Tier 3 ($149/node/mo) for the live in-cluster topology + shadow AI/ML workload discovery + Zero-Knowledge runtime detection no other commercial agent ships end-to-end.