Эта страница ещё не переведена: показана английская версия.

Microsoft Azure¶

The health of an Azure subscription: resources that Azure reports as unavailable or degraded, and the Azure Monitor alerts that are fired and not closed.

DumbMonit reads two lists on management.azure.com, once a minute: Resource Health (Microsoft.ResourceHealth/availabilityStatuses, API 2022-10-01), the availability that the Azure platform announces for each resource, and the Azure Monitor alerts (Microsoft.AlertsManagement/alerts, API 2019-05-05-preview) that are fired and either new or acknowledged. The credential is a service principal: an Entra application with a client secret and the Reader role on the subscription. DumbMonit exchanges the secret for a one-hour token (client credentials flow, scope https://management.azure.com/.default) and keeps it until it expires. The Reader role can read and change nothing.

A stopped (deallocated) virtual machine is reported by Azure as unavailable because its owner asked for it: DumbMonit counts it apart and does not alert.

This is an Internet API: the address is the API's, not a machine's. DumbMonit needs outbound HTTPS to management.azure.com and login.microsoftonline.com; nothing inside your resources is contacted.

Validated against the vendor documentation only — not yet tested on a real system

This integration was built from Microsoft's REST API documentation, not against a live subscription. It only reads. Tell us what breaks.

Network ports to open¶

From the DumbMonit server (or its relay agent) to the device:

Port	Protocol	Purpose
443	TCP	Azure Resource Manager API (HTTPS, management.azure.com)
443	TCP	Microsoft sign-in, access token (HTTPS, login.microsoftonline.com)

What it watches¶

All metrics are prefixed dumbmonit_azure_.

Metric	Labels	Meaning
`resources`		Resources with a Resource Health status.
`resources_available`, `resources_degraded`, `resources_unavailable`, `resources_stopped`, `resources_unknown`		Resources by state; `stopped` is unavailable on the owner's request.
`resource_unhealthy`	`resource`, `type`, `resource_group`, `state`	Value 1 for each degraded or unavailable resource.
`alerts_readable`		0 when the alerts cannot be read (see Troubleshooting); Resource Health is still reported.
`alerts_active`	`severity` (`Sev0` to `Sev4`)	Fired alerts, new or acknowledged, by severity.
`alerts_high_active`		Fired alerts of severity Sev0 or Sev1.
`alerts_total`		All fired alerts.

At most 200 named resources and 200 named alerts are listed; the counts stay exact. Lists are read over 20 pages at most.

The built-in rules that apply:

Azure resource unavailable: a resource unavailable for 10 minutes (Critical).
Azure resource degraded: a resource degraded for 15 minutes (Warning).
Azure Monitor alert Sev0 or Sev1: a Sev0 or Sev1 alert fired and not closed for 5 minutes (Warning).

The monitor page¶

The Azure panel says what is wrong in a sentence, shows the number of resources, the unavailable and degraded ones and the active alerts, then lists the unhealthy resources (and the stopped ones) and the active alerts.

Create a read-only application for DumbMonit¶

In the Azure portal, open Microsoft Entra ID > App registrations > New registration, name it DumbMonit, keep single tenant, then Register.
On the Overview page, copy the Application (client) ID and the Directory (tenant) ID.
Open Certificates & secrets > New client secret, choose a lifetime, then copy the secret Value: it is shown once.
Open Subscriptions > your subscription > Access control (IAM) > Add > Add role assignment, pick the Reader role, select the DumbMonit application, then Review + assign.
Copy the Subscription ID from the subscription Overview page.
In DumbMonit, enter management.azure.com as address, paste the client ID and secret, and fill the Tenant ID and Subscription ID options.

Required permissions¶

Reader (role on the subscription) — Reads Resource Health and the active Azure Monitor alerts; changes nothing.

Warning

A stopped or deallocated virtual machine is reported by Azure as unavailable on the user's request: it is counted apart and raises no alert. Azure Monitor alerts need the Microsoft.AlertsManagement resource provider registered on the subscription; without it only Resource Health is monitored.

Vendor documentation: https://learn.microsoft.com/rest/api/resourcehealth/availability-statuses/list-by-subscription-id and https://learn.microsoft.com/rest/api/monitor/alertsmanagement/alerts

Credentials¶

Credential	Fields
Entra application (service principal)	Application (client) ID and client secret Value.

Address: management.azure.com. Options:

Tenant ID (required): the Directory (tenant) ID of the application.
Subscription ID (required): the subscription to watch.
Sign-in URL: https://login.microsoftonline.com by default; change it only for a sovereign cloud.
Timeout per request (15 s).

The client secret expires (at most 24 months): renew it before, or monitor it with the Microsoft Entra ID integration.

Troubleshooting¶

"Microsoft refused the application": the client ID or secret is wrong, or the secret has expired.

"Azure refused to read the Resource Health (403)": the application does not have the Reader role on that subscription; role assignments can take a few minutes to apply.

"Azure knows no subscription …": check the Subscription ID and the Tenant ID.

alerts_readable is 0: the application cannot read Azure Monitor alerts, or the Microsoft.AlertsManagement resource provider is not registered on the subscription (Subscriptions > Resource providers).