How to Monitor Non‑Kubernetes ECS Apps with Alibaba Cloud Managed Prometheus

Introduction

Prometheus + Grafana is the de‑facto standard for observability in cloud‑native environments. While it works well for Kubernetes, many legacy systems still run on non‑cloud‑native ECS instances, requiring a similar metric‑collection loop.

Typical ECS Deployment Scenarios

Scenario 1 – Pure Public‑cloud VPC

Applications and supporting components (databases, middleware) are deployed on multiple ECS instances inside one or more VPCs. Metrics from the OS, components and business services need to be collected.

Scenario 2 – Public‑cloud VPC + On‑premise IDC

In addition to VPC‑based ECS, the workload communicates with an on‑premise IDC via dedicated lines. A unified metric platform must cover both sides.

Scenario 3 – Public‑cloud VPC + Multi‑cloud ECS

Besides Alibaba Cloud VPC, the service also talks to ECS instances in other clouds. A global view of all metrics is required.

Why Self‑Built Prometheus Is Painful

Multiple isolated VPCs force repeated independent Prometheus deployments, raising operational cost.

Each stack requires installing and configuring Prometheus, Grafana, Alertmanager and exporters, which is time‑consuming.

Lack of native Alibaba Cloud ECS service‑discovery; implementing it needs custom Go code that integrates with the ECS POP API, complicating upgrades.

Open‑source Grafana dashboards are often insufficiently customized for production needs.

Alert templates for common components are missing, forcing users to create them from scratch.

Alibaba Cloud Managed Prometheus Capabilities

Alibaba Cloud Prometheus is a fully managed, hybrid‑cloud Prometheus service that integrates with open‑source ecosystems, provides out‑of‑the‑box dashboards, and supports multi‑instance aggregation, unified Grafana data source, and unified alerting.

Metric collection for ECS applications is offered via three methods:

Managed exporters : Pre‑deployed exporters for MySQL, Redis, etc. Users only configure target IP/port in the console; metrics are collected from both VPC and IDC.

Custom (unmanaged) exporters : Deploy user‑written exporters for unsupported components, then register them via ServiceDiscovery in the console.

Node/Windows exporters : Deploy on each ECS to gather OS‑level metrics; Node exporter is supported now, Windows exporter is forthcoming.

Self‑Built vs Managed: Feature Comparison

A side‑by‑side comparison highlights differences in deployment effort, maintenance, feature set, and cost.

Step‑by‑Step Guide to Monitor ECS with Alibaba Cloud Prometheus

Step 1 – Create a VPC Instance

Log in to the Prometheus console, create a new instance, and specify the VPC, VSwitch, SecurityGroup and Grafana workspace.

Step 2 – Add Component Monitoring

Supported components include Node exporter, MySQL, Redis, Elasticsearch, Kafka, Nginx, MongoDB, PostgreSQL, RabbitMQ, RocketMQ, BlackBox, etc. The service supports static_configs and aliyun_sd_configs for service discovery.

For MySQL, create a new MySQL integration, fill in name, address, port, credentials, and save.

Step 3 – View Dashboards

The managed Grafana instance is automatically linked; click the installed exporter icon in the integration center to open the corresponding dashboard.

Step 4 – Configure Alerts

In the integration center, open the alert page for a component (e.g., MySQL) and create Prometheus alert rules.

Conclusion

Alibaba Cloud Managed Prometheus provides a zero‑maintenance, out‑of‑the‑box solution for observing OS, middleware and business‑level metrics on ECS, covering both cloud‑native and legacy environments. Ongoing updates will add more exporters such as Windows, JMX, ClickHouse, Jenkins, and Process.

Reference links: https://help.aliyun.com/document_detail/122123.html, https://help.aliyun.com/document_detail/251830.html, https://help.aliyun.com/document_detail/161838.html, https://help.aliyun.com/document_detail/427600.html, https://help.aliyun.com/document_detail/331981.html