You were redirected from a different version of the documentation. Click here to go back.

FAQ: Monitoring and Alerts

This addresses common questions about Ops Manager and how it monitors MongoDB instances and alerts you of issues with those deployments.

Host Configuration

How do I add a new host or server?

See Add Existing MongoDB Processes to Ops Manager.

Can Ops Manager monitor itself?

Yes. You can use Ops Manager to monitor the backing MongoDB instances that hold data for Ops Manager and the Backup Daemon. Create a separate project for the backing instances.

You cannot, however, use Ops Manager to backup or automate the backing instances.

Ops Manager also provides built-in basic monitoring of all backing databases through system alerts.

Can I monitor Kerberos-enabled instances?

Yes. Ops Manager does support monitoring for Kerberos-enabled MongoDB instances. To learn how this works, see Configure the MongoDB Agent for Kerberos.

Monitoring Agent

Do I need a Monitoring Agent for every MongoDB instance?

No. In your Ops Manager project, a single Monitoring Agent connects to all MongoDB databases. Configure firewalls to allow the Monitoring Agent to connect across data centers and servers.

Beginning with Monitoring Agent version 5.0.0, you can run multiple agents to distribute monitoring assignments and provide failover. Ops Manager distributes monitoring assignments among up to 100 running Monitoring Agents. If you run more than 100 Monitoring Agents, the additional agents behave as “standby” agents.

Where should I run the Monitoring Agent?

The amount of resources the Monitoring Agent requires varies depending on infrastructure size, the number of servers and the databases it is monitoring. Run the agent on an existing machine with additional capacity that does not run a mongod instance. You may also run the Monitoring Agent on a smaller dedicated instance.

The Monitoring Agent load scales with the number of monitored mongod plus mongos processes and the number of databases in your MongoDB environment.

For production environments, it is recommended to install the Monitoring Agent on a dedicated server, and not on the the same host as a data bearing mongod instance. This allows you to perform maintenance on the mongod and its host without affecting the monitoring for your deployment. Additionally, a Monitoring Agent may contend for resources with the mongod.

You can install the Monitoring Agent on the same system as an arbiter, a mongos, or an application server depending on the requirements of these services and available resources.

Can I run the Monitoring Agent on an AWS micro server?

If you monitor five or fewer mongod instances, you can use an AWS micro server.

Why can’t the Monitoring Agent connect to my host?

The most common problem is that the agent is unable to resolve the hostname of the host. Check DNS and the /etc/hosts file.

The second most common problem is that there are firewall rules in place that prohibit access to the host from the agent.

To test the connection, login to the host running the agent and try to connect using mongosh:

mongosh <hostname>:<port>/test


Ops Manager does not support port forwarding.

Why does the Monitoring Agent connect with hostnames instead of IP addresses?

By default, the Monitoring Agent resolves hostnames to connect. If the agent cannot connect by resolving a hostname, you can force the Monitoring Agent to prefer an IP address over its corresponding hostname for a specific IP address. Preferred hostnames also allow you to specify the hostname to use for servers with multiple aliases. This prevents servers from appearing multiple times under different names in the Ops Manager interface.

To create a preferred hostname, go to Project Settings and add a Preferred Hostnames entry. For details, see Edit Project Settings

How do I setup and configure the agent?

See the README file included in the agent download.

How do I delete a Monitoring Agent from Ops Manager?

See Remove Monitoring Agents from Ops Manager.

Data Presentation

What are all those vertical bars in my charts?

A red bar indicates a server restart.

A orange bar indicates the server is now a primary.

A brown bar indicates the server is now a secondary.

Data Retention

What is the data retention policy for Ops Manager?

Ops Manager retains two distinct types of data: metrics, which describe usage; and snapshots, which back up your data.

Ops Manager preserves:

  • Ops Manager preserves metric data according to Ops Manager’s Monitoring Data Retention settings. Ops Manager administrators can modify the settings by selecting Admin, then General, then Ops Manager Config, then the Miscellaneous tab, and then scrolling to Default Monitoring Data Retention.
  • Snapshots according to their retention policy.


What do the alert conditions mean?

For a reference on the alert conditions, see Review Alert Conditions.

What alerts are configured by default?

See Configure Alert Settings for the default alert configurations as well as steps to add new alerts or modify existing alerts, including modifying the alert frequency.