Atlas Built‑in Monitoring
The Atlas platform supplies native dashboards that expose CPU, memory, and disk usage in real time. Its Performance Advisor highlights slow queries and suggests index changes, making early detection possible.
Alert rules can be defined with threshold values for replication lag, connection spikes, and IO pressure, delivering email or webhook notifications. The configuration UI keeps the process simple and avoids extra agents.
- Log into Atlas and navigate to the Monitoring tab.
- Select the cluster and enable the Performance Advisor.
- Define alert thresholds for CPU, replication lag, and disk latency.
- Choose notification channels and save the policy.
Percona Monitoring and Management (PMM)
PMM combines Grafana visualizations with VictoriaMetrics storage to render query trends, cache hits, and lock contention. Its MongoDB exporter pulls metrics directly from the server, preserving granularity.
For teams that self‑host, PMM offers a Docker compose bundle that deploys all components with a single command, reducing manual setup errors. The Query Analytics panel surfaces execution plans and index usage, enabling targeted tuning.
- Download the PMM‑Server Docker image.
- Run the docker‑compose file to start Grafana, Metrics, and the exporter.
- Register the MongoDB instance via the PMM client.
- Access the Grafana UI and import the MongoDB dashboard template.
Prometheus with MongoDB Exporter
Prometheus scrapes metrics from the official MongoDB exporter, storing them as time‑series data that can be queried with PromQL. This approach works with any environment, from bare metal to Kubernetes.
The exporter provides latency, oplog size, and replication lag counters, which can be visualized in Grafana dashboards built from community templates. Alert rules are expressed as Prometheus alerts, allowing integration with Alertmanager.
- Deploy the MongoDB exporter as a sidecar or separate pod.
- Configure Prometheus to scrape the exporter endpoint.
- Create alerting rules for replication lag and high operation latency.
- Link Alertmanager to your incident response channel.
Datadog MongoDB Integration
Datadog collects host metrics, process statistics, and MongoDB specific counters via an agent that runs on each node. The integration supplies pre‑built dashboards that show query performance, cache efficiency, and shard health.
Alerting in Datadog can be scoped by tags, enabling fine‑grained policies for each replica set member. The platform also supports anomaly detection, which flags out‑of‑range values without manual thresholds.
- Install the Datadog agent on every MongoDB server.
- Enable the MongoDB integration in the agent configuration file.
- Restart the agent to begin metric collection.
- Apply dashboard templates and configure tag‑based alerts.
New Relic MongoDB Observability
New Relic ingests query traces, connection counts, and storage metrics through its MongoDB plugin, presenting them in a unified UI alongside application data. The service highlights slow operations and index usage patterns.
Its alerting engine supports NRQL queries, allowing custom conditions such as average query time exceeds 200 ms for five minutes. The platform also offers a service map that visualizes dependencies between services and databases.
- Enable the MongoDB plugin in the New Relic agent.
- Provide connection strings for each cluster.
- Verify data flow in the New Relic dashboard.
- Write NRQL alerts for latency and replication lag thresholds.