- 3 minutes to read

Monitoring Cosmos DB in Azure

Monitor your Azure Cosmos DB estate from Nodinite with the same operational visibility you already use for the rest of your integration landscape. This page explains how the Nodinite Azure agent discovers SQL (Core) API databases and containers, evaluates their availability, and surfaces changes in a regular Monitor View.

New 7.9.x This feature adds native monitoring for Azure Cosmos DB databases and containers in Nodinite, including automatic discovery, availability monitoring, and configuration visibility.

  • ✅ Discover databases and containers automatically without manual lists
  • ✅ Surface reachability problems as visible resources instead of silently skipping them
  • ✅ Keep inventory current as databases or containers are added or removed in Azure
  • ✅ Show configuration facts such as RU/s, autoscale maxima, partition keys, and TTL

Nodinite monitors each Azure Cosmos DB account through the configured connection details. Once the account is known, the agent discovers all databases and all containers beneath them and presents them as monitored Resources. This makes Cosmos DB availability and drift visible in the same workflow as other Azure services.

graph LR subgraph "Azure" A[Cosmos DB Account] B[Database] C[Container] end subgraph "Nodinite" D[Monitoring Agent] E[Monitor View Resource] end A --> B B --> C D --> A D --> E

Diagram: The Azure agent checks the Cosmos DB account and exposes each database and container as a monitored resource in Nodinite.

Monitoring Features

  • Automatic discovery — New databases and containers appear without manual reconfiguration.
  • Continuous re-discovery — Each polling cycle updates the inventory and removes entities that are genuinely gone.
  • Category-based monitoring — Databases and containers are grouped in Categories for clear ownership and filtering.
  • Role-based visibility — Share access through one or more Role-based Monitor Views.
  • Configuration facts — The resource shows RU/s, autoscale max, partition key, and TTL details.

State Evaluation for Cosmos DB

Nodinite creates a Resource for each monitored database and container. The resource state is based on the agent's ability to read it successfully.

State Meaning What the user sees
OK The database or container is reachable and readable Green state with no warning text
Unavailable The account is unreachable, the connection string is invalid, or the container cannot be read A red resource with an explanatory message

The monitoring model is intentionally simple:

  • A read success is OK — healthy resources remain green.
  • A read failure is visible — a broken account or container becomes UNAVAILABLE instead of disappearing silently.
  • Resources are removed only after sustained absence — if a database or container is genuinely deleted in Azure, it disappears after three missed cycles.

Databases

Each Cosmos DB database is represented as a resource with details such as:

  • Account name
  • Number of containers
  • Provisioned RU/s
  • Autoscale max RU/s

If the account becomes unreachable or the database cannot be read, the resource is marked Unavailable and the user sees the connection or access problem directly in the monitor view.

Containers

Each container resource shows the container's own configuration facts:

  • Partition key path
  • Provisioned RU/s
  • Autoscale max RU/s
  • Default TTL

If a single container cannot be read, only that container becomes Unavailable. Healthy containers remain visible and usable.

Alert History and Operational Awareness

When an incident happens, you can review the event history for the affected database or container in the same way as other monitored resources. Use the history capabilities of your Monitor View to review whether the issue started with a connection problem, an access issue, or a resource removal.

Frequently Asked Questions

What is monitored?

Nodinite monitors Azure Cosmos DB SQL (Core) API databases and containers, including reachability, inventory drift, and configuration facts. It does not provide threshold-based numeric alerting for RU consumption or latency metrics.

What happens if the account is unreachable?

The monitored account resource becomes Unavailable with a plain-language explanation. This makes connection-string problems, rotated keys, and access issues visible immediately in Nodinite.


Next Step