On this page
Atlas collects and displays metrics for your servers, databases, and MongoDB processes.
Monitor database deployment metrics to identify performance issues and determine whether your current database deployment meets your requirements. For more information on the metrics available to monitor your database deployments, see Review Available Metrics.
The number of servers that Atlas displays on the
Metrics page at any given time depends on the browser
screen size. Use the Toggle Members section to control
which servers Atlas displays. Hover over the
icons to find out which servers they represent.
You can view metrics in the following ways:
- Project Overview
- The Overview tab displays all the database deployments in an Atlas project and features core metrics per database deployment.
- Atlas Serverless Instances
- View the metrics for a specific serverless instance in an Atlas project.
- Atlas Replica Sets
- View the metrics for a specific replica set in an Atlas project.
- Atlas Sharded Clusters
- View the metrics for a specific sharded cluster in an Atlas project.
- MongoDB Processes
- View the metrics for a specific MongoDB process in an Atlas cluster.
- Real-Time Performance
- View real-time performance metrics for a specific Atlas database deployment in a project.
- Atlas Search
- View Atlas Search metrics for Atlas clusters with at least one active Atlas Search index.
You can monitor the following metrics to quickly gauge the health of your database deployment.
Number that indicates the total bytes of used disk space for the cluster.
Monitor the combined size of your data and MongoDB operational data (buffer, journal, and log files) on the cluster.
To learn more, see Fix Storage Issues.
Number that indicates the efficiency of read operations run on MongoDB.
Monitor query targeting metrics to identify inefficent queries.
To learn more, see Fix Query Issues.
Normalized System CPU
Percentage of time that requests are issued to and serviced by disk. This metric includes requests from any process, not just MongoDB processes.
Monitor whether utilization is high. Determine whether to increase the provisioned IOPS or upgrade the cluster.
To learn more, see Fix IOPS Issues.
Atlas stores metrics data at increasing granularity levels. For each increasing granularity level, Atlas computes the metrics data based on the averages from the previous granularity level. The length of retention depends on the granularity. Atlas computes the metrics data based on the averages from the previous granularity level.
Atlas gathers metrics data at a 1-minute granularity unless you qualify for premium monitoring.
After 48 hours' worth of data is collected, each group of 60 minutes is compacted into a single unit of an hour. After 63 days, each group of 24 hours is compacted into a single unit of a day.
If you have at least one cluster that's
M40 or larger, Atlas
automatically enables premium monitoring for all clusters in the project. With
premium monitoring enabled, Atlas gathers metrics data at a 10-second
granularity. Premium monitoring remains enabled for all clusters in the project until you
downgrade or terminate your last
Premium monitoring will become available in stages. If you have an
cluster or larger but don't see the premium monitoring granularity yet, you
should see it soon.
Atlas retains metrics data for a period of time that depends on the granularity of the data:
Duration of Retention
Premium monitoring only.
Atlas retains all database-specific statistics. MongoDB log data is retained at a maximum rate of 2000 lines per 2 minutes.
M0free clusters and
M2/M5shared clusters support a subset of the metrics and charts available. For complete documentation on the limitations of
M0/M2/M5clusters, see Atlas M0 (Free Cluster), M2, and M5 Limitations.
- Atlas pauses monitoring for
M0free clusters which have had no connection activity for 7 days. Monitoring resumes once a successful connection occurs through the Atlas Administration API, Driver,
mongosh, or Data Explorer.
- Serverless instances support a subset of the metrics and charts available. For complete documentation on the limitations of serverless instances, see Serverless Instance Limitations.