Managed Service for Prometheus:Monitor a Windows ECS instance

How it works

1. Windows Exporter runs on the ECS instance and exposes system metrics on an HTTP port (default: 9182).

2. The Windows component in ARMS discovers the instance and scrapes metrics at a configured interval.

3. Pre-built dashboards in the ARMS console visualize CPU, memory, disk, and network data.

Prerequisites

Make sure that:

• A Windows Server ECS instance is deployed in a VPC.

• You have administrator access to the Windows instance.

• Access to the ARMS console

Step 1: (Optional) Install Windows Exporter

ARMS supports two deployment methods. To use automatic installation, skip to Step 2. The system installs and configures the exporter based on service discovery rules.

To install Windows Exporter manually:

1. Log on to the ECS instance as an administrator.

2. Download the Windows Exporter installation package (windows_exporter-1.22.0-amd64.msi).

3. Double-click the .msi file or run the following command. Available MSI installer parameters: Example with a custom port and specific collectors:

   Parameter	Description	Default value
   ENABLED_COLLECTORS	Comma-separated list of metric collectors to enable	cpu, cpu_info, memory, process, tcp, cs, logical_disk, net, os, system, textfile, time
   LISTEN_PORT	Port that the exporter listens on	9182
   LISTEN_ADDR	IP address to bind to	Empty (all local addresses)
   METRICS_PATH	HTTP path at which metrics are served	/metrics
   TEXTFILE_DIRS	Directories containing custom textfile collector files	None
   CONFIG_FILE	Path to a YAML configuration file	None
   Installation directory	--	C:\Program Files (x86)\windows_exporter

      msiexec /i windows_exporter-1.22.0-amd64.msi

      msiexec /i windows_exporter-1.22.0-amd64.msi LISTEN_PORT="9183" ENABLED_COLLECTORS="cpu,memory,logical_disk,net,os,system"

4. Verify the exporter is running. Open a browser or run the following command and confirm that Prometheus-format metrics are returned: Expected output: You can also check Windows Console (Services) to confirm the windows_exporter service is Running.

      curl localhost:9182/metrics

      # HELP windows_os_info OperatingSystem.Caption, OperatingSystem.Version
      # TYPE windows_os_info gauge
      windows_os_info{product="Windows Server 2019 Datacenter",version="10.0.17763"} 1
      # HELP windows_cpu_time_total Time that processor spent in different modes (dpc, idle, interrupt, privileged, user)
      # TYPE windows_cpu_time_total counter
      windows_cpu_time_total{core="0,0",mode="dpc"} 4.421875
      ...

5. (Optional) Edit the config.yml file to change the port or collectors. The Windows Exporter documentation covers all configuration options. Restart the service after saving:

      sc stop windows_exporter && sc start windows_exporter

Step 2: Enable the Windows component

1. Log on to the Managed Service for Prometheus console.

2. In the left-side navigation pane, click Integration Center.

3. Search for Windows and click it.

   

4. On the Start Integration tab, configure the following parameters and click OK.

   Parameter	Description
   Select the environment type	Valid value: ECS(VPC).
   WindowsExporter Installation Mode	Select Automatic Installation to let ARMS deploy the exporter, or Self Installation if you completed Step 1.
   Host Service Discovery Mode	Options: Unconditional (Select all Windows instances in the current VPC), Stain Label Selection, Tag Label Selection, IP CIDR, or Instance ID.
   Collect TextFile	Absolute path to a textfile collector directory. Environment variables are not supported. Example: /home/app/textfiles.
   Windows-Exporter Service port	Port that Windows Exporter listens on. Pre-filled with 9182.
   Security Groups Are Automatically Configured	ARMS verifies that the exporter port is open in the ECS security group and adds inbound rules if needed.
   Custom ECS Tag Injection	ECS tag keys to inject into scraped metrics as labels.
   Metric Scrape interval (unit/second)	Scrape interval. Default: 15 seconds.

   

Step 4: Configure alerting

Set up alert rules to get notified when metrics such as CPU usage, available memory, or disk space cross thresholds.

Create an alert rule for a Prometheus instance.

Default metric collectors

The following collectors are enabled by default.

To add collectors, edit config.yml and restart the service. The Windows Exporter documentation lists all available collectors.