SYSTEM Cited by 2 sources
Amazon Managed Grafana¶
Amazon Managed Grafana (AMG) is AWS's fully managed Grafana service — the Grafana visualization engine operated by AWS, with native integrations to AWS data sources (CloudWatch, AMP / Managed Prometheus, Timestream, X-Ray, OpenSearch), authentication via AWS SSO / IAM Identity Center, and per-workspace isolation. Stub page — expand on future AMG-internals sources.
Why it shows up on this wiki¶
AMG is the customer-facing dashboarding tier that pairs with the AWS observability stack:
- CloudWatch / CloudWatch Logs Insights = metrics + logs source.
- Amazon Managed Prometheus = Prometheus-API-compatible metrics source.
- X-Ray = traces source.
- AMG = the dashboard / alert / visualisation surface over all of the above.
AWS operates the Grafana server, including upgrades, HA, and authentication integration; customers get per-workspace role-based access and per-workspace plugin / data-source configuration.
Tenancy shape seen in ingested sources¶
At Generali Malaysia, AMG is used to give each application owner per-project dashboards scoped to an EKS namespace. One AMG workspace exposes many dashboards; dashboards are sliced by the cluster / namespace dimension that already segments the K8s workload. This is the same tenancy axis as the cost-allocation-tag scheme (eks-cost-allocation-tags) — one consistent dimension across billing and observability.
Dashboards shown:
- Cluster health
- Node performance
- Pod resource utilization
- Application performance indicators (custom per-project)
All driven from CloudWatch as data source via the CloudWatch → AMG integration.
Seen in¶
- sources/2026-03-23-aws-generali-malaysia-eks-auto-mode — the per-namespace observability substrate for Generali's multi-tenant EKS cluster. One AMG workspace, N dashboards (one per business- unit project). CloudWatch as data source, not direct Prometheus.
-
sources/2026-04-06-aws-unlock-efficient-model-deployment-simplified-inference-operator-setup-on-amazon-sagemaker-hyperpod — the dashboard surface for LLM-inference observability on AWS's 2026-04-06 SageMaker HyperPod Inference Operator. Post cites "built-in integration with HyperPod Observability provides immediate visibility into inference metrics, cache performance, and routing efficiency through Amazon Managed Grafana dashboards." Key inference metrics surfaced via AMG: time-to-first-token (TTFT), end-to-end inference latency, GPU utilization. No internal architecture disclosed — AMG appears here as the dashboard layer, not the metrics pipeline.
-
sources/2026-08-13-aws-adobe-firefly-simplified-observability-with-amazon-managed-prometheus — Adobe Firefly uses AMP as the managed high-cardinality metrics backend for EKS GPU-training telemetry; AMG is the article's named native dashboard integration. The source does not disclose its dashboards, alerting rules, or the exact AMG workspace configuration. (Source: sources/2026-08-13-aws-adobe-firefly-simplified-observability-with-amazon-managed-prometheus)
Related¶
- systems/grafana — the OSS project AMG is based on.
- systems/amazon-managed-prometheus — the AWS-native Prometheus- compatible metrics backend that AMG pairs with (when you want Prometheus-format ingestion rather than CloudWatch).
- systems/aws-eks — most common AMG customer context.
- concepts/observability