Use the EDB observability stack

This how-to guide explains how to use the EDB observability stack—Prometheus, Grafana, and Loki—as deployed with Hybrid Manager (HM) on Kubernetes.

Important: This guidance is specific to Hybrid Manager deployments. When using Hybrid Manager, the observability stack is deployed and managed as part of the platform toolchain. This is the recommended and supported mechanism for monitoring Hybrid Manager components and EDB-managed Postgres services.

For architectural guidance, see Monitoring Kubernetes best practices. For troubleshooting observability issues, see Troubleshooting Kubernetes.

What is included

Hybrid Manager deployments typically include:

• Prometheus → metrics collection for:
• Kubernetes cluster components
• Hybrid Manager components (UI/API/Operators)
• EDB-managed Postgres clusters
• Grafana → dashboards for visualizing metrics
• Loki → log aggregation
• Fluent Bit → log forwarding from Kubernetes Pods to Loki

These components are deployed and configured automatically by Hybrid Manager.

What is not included

The following are outside the scope of the standard EDB observability stack:

• External log forwarding (Splunk, ELK, etc.) → customer-managed and
• External metrics export (Datadog, CloudWatch, etc.) → customer-managed
• Application-specific dashboards beyond EDB Postgres and Hybrid Manager components → customer can extend Grafana as needed

Accessing Prometheus

Prometheus is deployed in the monitoring namespace (default).

To access Prometheus:

• Use the configured ingress or LoadBalancer endpoint (if exposed).
• Or use kubectl port-forward to access Prometheus UI locally.

Prometheus provides:

• Metrics explorer
• Time series graphs
• Raw metrics for troubleshooting

Using Grafana dashboards

Grafana is deployed in the monitoring namespace (default).

Grafana includes pre-built dashboards for:

• Kubernetes cluster health
• Node resource usage
• Pod resource usage
• Hybrid Manager components (UI/API/Operators)
• EDB-managed Postgres clusters (database metrics, replication, storage, connections)

To access Grafana:

• Use the configured ingress or LoadBalancer endpoint (if exposed).
• Or use kubectl port-forward to access Grafana locally.

Common tasks:

• View node and cluster health
• Monitor Postgres replication and performance
• Correlate Hybrid Manager component behavior with cluster events

Accessing logs with Loki

Loki is deployed in the monitoring namespace (default).

Logs are forwarded to Loki using Fluent Bit.

Logs typically available:

• Hybrid Manager component logs (UI/API/Operators)
• Postgres cluster logs (container stdout/stderr → includes Postgres server logs)
• Kubernetes system component logs (if configured)

To access Loki:

• Use the Grafana "Explore" tab.
• Or use Grafana dashboards with log panels.

Common tasks:

• Search Hybrid Manager component logs
• Troubleshoot Postgres errors
• Correlate log events with metrics

Typical monitoring workflow

Recommended workflow:

1. Monitor dashboards → Grafana
2. Investigate metrics → Prometheus / Grafana Explore
3. Correlate with logs → Loki via Grafana
4. Use Kubernetes CLI tools as needed for deeper investigation

Summary checklist

• Use Prometheus to explore and validate cluster and application metrics.
• Use Grafana dashboards to monitor Hybrid Manager and Postgres health.
• Use Loki to search and analyze logs for Hybrid Manager components and Postgres.
• Extend dashboards as needed for environment-specific needs.
• Understand what is included and what requires customer tooling.

Related topics

• Monitoring Kubernetes best practices
• Troubleshooting Kubernetes
• Manage Kubernetes networking
• Manage Kubernetes storage
• Kubernetes in Hybrid Manager