Edge Node Observability Stack#

The Edge Node Observability stack is intended for Edge & Host Managers and Host Operators to monitor the telemetry from the host level of the Edge Node, including provisioning. Refer to Hosts (Edge Node) for more details.

Key components of the observability stack that allow collection, storage and consumption of metrics, logs, and traces are:

  • Grafana UI - provides the frontend for visualization.

  • OpenTelemetry Collector (Customized):

    • Provides logs collection pipelines with log transformation.

    • Collects supplementary metrics from the Edge Orchestrator services.

    • Routes logs and metrics to appropriate tenants based on metadata.

  • Grafana Mimir - used as metrics backend.

  • Grafana Loki - used as logs backend.

Note

To restrict access to data associated with individual tenants, a grafana-proxy custom extension is used while accessing datasources. Refer to :doc: ../concepts/multitenancy for more details on how multitenacy in observability is hanlded.

System Context diagram for Edge Node Observability Stack

Deployment#

The stack is separated into the following charts

  • edgenode-observability - provides core services for the observability stack (for example, Grafana UI, Grafana Mimir, Grafana Loki, OpenTelemetry Collector).

  • edgenode-dashboards - provides Grafana UI dashboards for Edge Node Hosts.

Storage#

Grafana Mimir and Grafana Loki are configured to use:

  • AWS S3 for cloud-based deployments

  • S3-compatible storage (MinIO) for on-prem deployments