# Datadog Integration > Source-anchored research for Datadog Integration. Run status: COMPLETED. Parent: [MongoDB Monitoring and Observability](https://llms-explorer.com/tree/mongodb-monitoring-and-observability/) · 3 facets · 18 facts · page: https://llms-explorer.com/tree/datadog-integration/ ## Problems, failure modes and limitations - The Agent MongoDB check supports custom find, count and aggregate metrics. Database Monitoring adds query snapshots, slow-query metrics, execution plans and collection insights beyond the standard integration. (low confidence) (verified as of 2026-10-23) — [source](https://github.com/DataDog/integrations-core/blob/master/mongo/README.md) - Datadog's standard MongoDB integration distinguishes Database Monitoring as an additional product for query snapshots, slow-query metrics, database load and execution plans. (low confidence) (verified as of 2026-10-23) — [source](https://github.com/DataDog/integrations-core/blob/master/mongo/README.md) ## Comparisons and alternatives - Atlas provides cluster, replica-set and database tags on relevant metrics. Custom tags require opt-in and apply to metrics rather than alerts. (low confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/tutorial/datadog-integration.md) - Atlas warns that mongodb.atlas.replstatus.health can remain healthy when nodes stop sending pings, or become inaccurate during partitions and Datadog ingestion problems. Configure no-data monitoring rather than treating health=1 alone as proof of availability. (low confidence) — [source](https://www.mongodb.com/docs/atlas/tutorial/datadog-integration.md) - Atlas exposes organization, project, cluster and database tags. User-defined resource tags are optional and apply to metrics rather than alerts. (medium confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/tutorial/datadog-integration.md) ## Facts and statements - Atlas pushes metrics into Datadog; setup uses a Datadog API key in Atlas project integrations. Both vendors document an M10+ cluster requirement. (high confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/tutorial/datadog-integration.md) - Atlas setup requires a Datadog API key. The CLI command atlas integrations create DATADOG creates or updates the project integration. (high confidence) — [source](https://www.mongodb.com/docs/atlas/tutorial/datadog-integration.md) - Atlas can deliver alerts as Datadog events through the same integration used for project metrics. (high confidence) — [source](https://www.mongodb.com/docs/atlas/tutorial/datadog-integration.md) - Atlas disables high-cardinality database, collection-latency and query-shape metric exports by default; enable sendDatabaseMetrics, sendCollectionLatencyMetrics and sendQueryShapeMetrics explicitly. (high confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/tutorial/datadog-integration.md) - The MongoDB check ships with the Datadog Agent. Its configuration uses mongo.d/conf.yaml and a monitoring user with clusterMonitor plus read permissions for the databases being monitored. (high confidence) (verified as of 2026-10-23) — [source](https://github.com/DataDog/integrations-core/blob/master/mongo/README.md) - MongoDB APM requires instrumentation in the application making database requests. OneUptime illustrates initializing Node.js dd-trace before importing the MongoDB client; installing the database metric check alone does not establish application tracing. (high confidence) — [source](https://github.com/DataDog/integrations-core/blob/master/mongo/README.md) - The Agent integration collects low-level metrics per mongod node, while higher-level collection/index statistics require topology-aware placement. Datadog documents separate replica-set and sharded-cluster layouts. (low confidence) — [source](https://github.com/DataDog/integrations-core/blob/master/mongo/README.md) - The MongoDB integration changelog records safeguards against monitoring overhead and failures: skipping explain on $mergeCursors pipelines to prevent potential crashes, and removing queryStats server-side sorting to avoid sort-memory failures. (low confidence) (verified as of 2026-07-08) — [source](https://github.com/DataDog/integrations-core/blob/master/mongo/CHANGELOG.md) - A February 2023 Terraform user reported INTEGRATION_FIELDS_INVALID and INVALID_ENUM_VALUE errors while configuring Datadog region values through the Atlas API. The captured thread records differing Atlas and Terraform documentation values and no follow-up confirming a fix. (low confidence) — [source](https://www.mongodb.com/community/forums/t/integration-fields-invalid-error-from-atlas-api-when-trying-to-enable-datadog-integration/214371) - A Datadog-authored Chef guest post shows automated Agent deployment and configuration of the MongoDB integration through Chef roles; this remains vendor-origin evidence. (low confidence) — [source](https://www.chef.io/blog/datadog-x-chef-how-and-why-these-two-platforms-go-together) - Atlas requires Project Owner access for this integration; an Organization Owner must first add themselves to the project with that role. (low confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/tutorial/datadog-integration.md) - Datadog recommends an Agent on each replica-set host, connecting to its local member. Low-level metrics require collection from every mongod, while higher-level collection/index statistics should be collected once. (low confidence) — [source](https://github.com/DataDog/integrations-core/blob/master/mongo/README.md) - MongoDB logs require a separate Agent configuration: enable logs_enabled in datadog.yaml and configure the MongoDB log source. Log collection is disabled by default. (low confidence) (verified as of 2026-10-23) — [source](https://github.com/DataDog/integrations-core/blob/master/mongo/README.md)