Infrastructure Observability

Overview

Infrastructure Observability in Cortx provides real-time visibility into the underlying systems that power your workspace. It enables authorized administrators to monitor compute resources, Kubernetes services, cloud infrastructure, billing metrics, and system health — all through connected dashboards.
This layer ensures platform reliability, operational transparency, and cost awareness.
Infrastructure Observability is primarily designed for governance, uptime monitoring, and operational control.

Who Can Access Infrastructure Observability?

Infrastructure Observability is restricted to:

  • Entity Administrators

  • Organization Administrators

  • Users with explicit infrastructure-level permissions

Regular users do not have access to this layer.
Access is governed by RBAC policies configured at the entity level.

What You Can Monitor

Infrastructure Observability provides visibility into:

  • Compute resource usage (CPU, memory, storage)

  • Kubernetes services and clusters

  • Node explorers

  • Shared database dashboards

  • Cloud infrastructure metrics

  • AWS billing dashboards

  • System-level uptime and performance indicators

These dashboards provide insight into what is running “under the hood” of the Cortx platform.

How It Works

Infrastructure Observability is integrated with Grafana dashboards.
Step 1: Navigate to Observability
  • Go to the Console section.

  • Select Observability.

  • Choose Infrastructure from the available layers.

Step 2: Select a Dashboard Widget

  • The Infrastructure view displays multiple widgets.

  • Each widget represents a connected monitoring dashboard (e.g., AWS Billing, Compute Resources).

Step 3: Drill Down
  • Click on any widget.

  • You will be routed to the corresponding Grafana dashboard.

  • Sign in via SSO (if prompted).

  • View detailed infrastructure metrics.

Dashboards support drill-down interactions for deeper inspection.

Key Capabilities

  • Dashboard-Level Visibility: View summarized infrastructure health through high-level widgets.
  • Drill-Down Analytics: Click any widget to access detailed Grafana dashboards.
  • Cost Monitoring: Track cloud billing metrics, including AWS billing dashboards.
  • System Health Monitoring: Monitor compute load, cluster health, and backend dependencies.

Best Practices

  • Regularly monitor billing dashboards to avoid cost overruns.

  • Review Kubernetes and compute dashboards during peak usage periods.

  • Investigate performance anomalies early to prevent service disruptions.

  • Ensure only authorized administrators retain infrastructure access.

  • Use Infrastructure Observability in conjunction with Application Observability for complete system insight.

Security & Access Control

Infrastructure Observability is restricted because it exposes backend operational data.

Only administrators should:

  • Access infrastructure dashboards

  • Analyze cloud billing

  • Review compute-level metrics

All access is controlled through RBAC policies.

Troubleshooting

Infrastructure Observability is restricted because it exposes backend operational data.

Only administrators should:

  • Access infrastructure dashboards

  • Analyze cloud billing

  • Review compute-level metrics

All access is controlled through RBAC policies.

Troubleshooting

Infrastructure Observability is integrated with Grafana dashboards.
I cannot see the Infrastructure layer.
You may not have administrator-level permissions. Contact your Entity Administrator.
Dashboard is not loading.
  • Ensure SSO authentication is active.

  • Verify Grafana access permissions.

  • Check network connectivity.

Billing dashboard data appears outdated.
Confirm that cloud integration sync is active and properly configured.

When to Use Infrastructure Observability

Use this layer when you need to:

  • Audit backend resource usage

  • Investigate performance issues

  • Monitor cloud costs

  • Review cluster and compute health

  • Validate system stability during high-load periods

Was this article helpful?

On this page