Migrate Self‑Hosted Prometheus + Thanos to Alibaba Cloud Managed Service

Background

Prometheus is the de‑facto standard for observability metrics. Many organizations still run self‑hosted stacks of open‑source Prometheus, Thanos, and Grafana for infrastructure and application monitoring.

Typical Self‑Hosted Deployment Scenarios

Kubernetes clusters – Alibaba Cloud Container Service, other cloud providers, or on‑premise IDC clusters.

Non‑Kubernetes hosts – Alibaba Cloud ECS, other cloud ECS instances, or VMs in a private IDC.

Long‑term storage is usually achieved by deploying one or more Thanos instances alongside each Prometheus instance.

Common Pain Points of Self‑Hosted Prometheus + Thanos

High operational effort to install exporters, configure dashboards and alert rules; community Grafana dashboards are often insufficient.

Multiple isolated deployments across clusters or VPCs increase cost and complexity.

Large‑scale clusters exceed the processing capacity of a single Prometheus instance.

Adding Thanos for long‑term storage introduces extra complexity and cost, especially in non‑Kubernetes environments that require Thanos Receiver.

Alibaba Cloud Managed Prometheus Capabilities

The service fully integrates with the open‑source Prometheus ecosystem, provides out‑of‑the‑box dashboards, built‑in alerting rules for common components (e.g., MySQL, Redis), and supports multi‑cloud / hybrid‑cloud monitoring with unified metric queries and Grafana data‑source integration.

Migration Process Overview

The migration is divided into three phases: metric collection, visualization, and alerting. The following steps apply to each deployment scenario.

Phase 1 – Metric Collection Migration

Configure Prometheus jobs to scrape targets and store data in the managed instance.

Alibaba Cloud Container Service (Kubernetes) Scenario

In the Access Center , select “Container Cluster Monitoring” and create a container environment instance.

In Access Management , add required components (e.g., Node Exporter, MySQL Exporter). Alibaba Cloud automatically generates scrape jobs, default dashboards, and alert rules.

Optionally create custom ServiceMonitor, PodMonitor, or Job resources.

Validate that the generated dashboards display correctly.

Self‑Hosted Kubernetes (Registered Cluster) Scenario

Register the self‑managed Kubernetes cluster as an Alibaba Cloud Container Service “registered cluster” and repeat the steps above.

Alibaba Cloud ECS Scenario

Create a “Host Monitoring” environment in the Access Center.

Deploy the appropriate exporters on each ECS host; Access Management generates the scrape configuration.

Optionally add custom collection rules.

Validate the default dashboards.

IDC (VM) Scenario

Connect the private IDC to an Alibaba Cloud VPC via a dedicated line.

Install Node and Process exporters on each VM.

Follow the same Access Center and Access Management steps as the ECS scenario.

Generic RemoteWrite Scenario

Create a “General Instance” in the Prometheus instance list.

Modify the self‑hosted Prometheus remote_write configuration to point to the Alibaba Cloud endpoint.

Verify ingestion in the Metrics Center.

Phase 2 – Visualization Migration

For container/Kubernetes and ECS scenarios, view built‑in dashboards via the “Dashboard Query” page in Access Management.

To use custom Grafana dashboards, create an “Expert‑edition Grafana” workspace, then either migrate an existing self‑managed Grafana or import specific dashboards.

In RemoteWrite scenarios, switch the Grafana data source to the corresponding Alibaba Cloud Prometheus instance.

Enable the global aggregation instance to achieve unified monitoring across multiple clusters or IDC.

Phase 3 – Alerting Migration

In Access Management, view default alert rules for each component under “Alert List”.

For custom alerts, create or import rules via the “Prometheus Alert Rules” or “Alert Rule Templates” menus.

Use the global aggregation instance to consolidate alerts across clusters.

References

https://prometheus.io/

https://help.aliyun.com/zh/prometheus/product-overview/product-introduction/

https://help.aliyun.com/zh/prometheus/use-cases/unified-multi-account-monitoring-based-on-prometheus-global-aggregation-instance

https://help.aliyun.com/zh/prometheus/container-observable

https://arms.console.aliyun.com/?#/intgr/integrations

https://arms.console.aliyun.com/#/intgr/integration-management

https://help.aliyun.com/zh/prometheus/user-guide/access-component-management

https://help.aliyun.com/zh/prometheus/user-guide/manage-container-environments-custom-collection-rules/

https://help.aliyun.com/zh/prometheus/use-cases/use-registered-kubernetes-clusters-to-implement-multi-cloud-kubernetes-cluster-monitoring

https://help.aliyun.com/zh/prometheus/host-observable

https://help.aliyun.com/zh/vpc/user-guide/connect-a-data-center-to-a-vpc

https://help.aliyun.com/zh/prometheus/user-guide/instance-management

https://help.aliyun.com/zh/prometheus/user-guide/write-open-source-prometheus-data-to-an-alibaba-cloud-prometheus-instance-through-a-remote-write-address

https://help.aliyun.com/zh/prometheus/user-guide/metrics-center/

https://help.aliyun.com/zh/prometheus/user-guide/search-out-of-the-box-with-a-large-tray

https://help.aliyun.com/zh/arms/observable-visualization-grafana-edition/manage-workspaces

https://help.aliyun.com/zh/arms/observable-visualization-grafana-edition/migrate-self-managed-grafana

https://help.aliyun.com/zh/arms/observable-visualization-grafana-edition/how-to-export-and-import-the-grafana-dashboard

https://help.aliyun.com/zh/prometheus/user-guide/data-query-and-grafana-data-source-settings

https://help.aliyun.com/zh/prometheus/user-guide/create-an-alert-rule-for-a-prometheus-instance

https://help.aliyun.com/zh/prometheus/use-cases/use-prometheusrule-to-create-prometheus-alert-rules