Integrate with a Monitoring Service

You can use a third-party application to view and analyze performance metrics that Atlas collects about your cluster.

At this time, you can either build a monitoring integration using the Atlas API or integrate Atlas with Datadog.

Build Monitoring Integrations with Atlas API

You can build a monitoring integration using the Atlas API monitoring and logs endpoints.

Integrate Atlas with Datadog

You can configure Atlas to send metric data about your project to your Datadog dashboards.


If your Atlas project is configured to send alerts and events to Datadog, you do not need to follow this procedure. Atlas sends project metrics to Datadog through the same integration used to send alerts and events.


Datadog integration is only available on M10+ clusters.

To integrate Atlas with Datadog, you must have a Datadog account and a Datadog API key. Datadog grants you an API key when you first create a Datadog account.

If you do not have an existing Datadog account, you can sign up at


To configure Atlas integration with Datadog:


Performance Metrics Available to Datadog

Datadog tracks the following metric data for your Atlas cluster. The metric names in parentheses are the names used in the Datadog UI. Each metric is measured once per minute.

Metric Name(s) Metric Type Description
CONNECTIONS (mongodb.atlas.connections.current) Process Number of open connections currently open on the cluster.
DB_STORAGE_TOTAL (mongodb.atlas.stats.totalstoragesize)
DB_DATA_SIZE_TOTAL (mongodb.atlas.stats.totaldatasize)
Process Total database storage size and data size on the cluster in bytes.
DOCUMENT_METRICS_RETURNED (mongodb.atlas.metrics.document.returned)
DOCUMENT_METRICS_INSERTED (mongodb.atlas.metrics.document.inserted)
DOCUMENT_METRICS_UPDATED (mongodb.atlas.metrics.document.updated)
DOCUMENT_METRICS_DELETED (mongodb.atlas.metrics.document.deleted)
Process Number of documents read or written per second.
OPCOUNTER_CMD (mongodb.atlas.opcounters.command)
OPCOUNTER_QUERY (mongodb.atlas.opcounters.query)
OPCOUNTER_UPDATE (mongodb.atlas.opcounters.update)
OPCOUNTER_DELETE (mongodb.atlas.opcounters.delete)
OPCOUNTER_GETMORE (mongodb.atlas.opcounters.getmore)
OPCOUNTER_INSERT (mongodb.atlas.opcounters.insert)
Process Number of operations per second, separated by operation type.
Process Average operation time in milliseconds, separated by operation type.
QUERY_TARGETING_SCANNED_OBJECTS_PER_RETURNED (mongodb.atlas.metrics.queryexecutor.scannedobjectsperreturned) Process Ratio measuring number of objects scanned over objects returned. Lower values indicate more efficient queries.
SYSTEM_NORMALIZED_CPU_KERNEL (mongodb.atlas.system.cpu.norm.kernel)
SYSTEM_NORMALIZED_CPU_IOWAIT (mongodb.atlas.system.cpu.norm.iowait)
SYSTEM_NORMALIZED_CPU_SOFTIRQ (mongodb.atlas.system.cpu.norm.softirq)
SYSTEM_NORMALIZED_CPU_GUEST (mongodb.atlas.system.cpu.norm.guest)
SYSTEM_NORMALIZED_CPU_STEAL (mongodb.atlas.system.cpu.norm.steal)
System Percent of time utilized by logical CPUs across various processes for the server. These values are normalized with respect to the number of logical CPU cores.
PROCESS_NORMALIZED_CPU_USER (mongodb.atlas.system.cpu.mongoprocess.user)
PROCESS_NORMALIZED_CPU_KERNEL (mongodb.atlas.system.cpu.mongoprocess.kernel)
PROCESS_NORMALIZED_CPU_CHILDREN_USER (mongodb.atlas.system.cpu.mongoprocess.childrenuser)
PROCESS_NORMALIZED_CPU_CHILDREN_KERNEL (mongodb.atlas.system.cpu.mongoprocess.childrenkernel)
Process Percent of time utilized by logical CPUs across various processes specific to the MongoDB process in the server. These values are normalized with respect to the number of logical CPU cores.
MEMORY_RESIDENT (mongodb.atlas.mem.resident)
MEMORY_VIRTUAL (mongodb.atlas.mem.virtual)
Process Memory (in MB) consumed by the MongoDB process on the server, separated by memory type.
OPCOUNTER_REPL_CMD (mongodb.atlas.opcountersrepl.command)
OPCOUNTER_REPL_UPDATE (mongodb.atlas.opcountersrepl.update)
OPCOUNTER_REPL_DELETE (mongodb.atlas.opcountersrepl.delete)
OPCOUNTER_REPL_INSERT (mongodb.atlas.opcountersrepl.insert)

Measure rate of database operations on MongoDB secondaries, as collected from the MongoDB serverStatus command’s opcountersRepl document.

You can view these metrics on the Opcounters - Repl chart, accessed via Cluster Metrics.

Process Measurements that apply to the MongoDB process’s oplog.
Disk Measure free disk space and used disk space (in bytes) on the disk partition used by MongoDB.
DISK_PARTITION_IOPS_WRITE (mongodb.atlas.system.disk.iops.writes)
Disk Measure throughput of IOPS for the disk partition used by MongoDB.