MongoDB Ops Manager and Cloud Manager
Parent: MongoDB Atlas · researched 2026-05-28T18:33:41.188Z· 42 sources · 24 concepts · skill mongodb-ops-manager
Ops Manager is MongoDB's self-hosted management platform for on-premises MongoDB deployments. Cloud Manager is the hosted SaaS equivalent (no infrastructure to manage). Both provide:
Overview
- Ops Manager is MongoDB's self-hosted management platform for on-premises MongoDB deployments. Cloud Manager is the hosted SaaS equivalent (no infrastructure to manage). Both provide: [source]
- Automation: Declarative goal-state management of MongoDB clusters (topology, version, config) [source]
- Monitoring: Real-time metrics, alerts, and dashboards [source]
- Backup: Continuous oplog-based backup with PITR [source]
App Database Sizing
- The App DB stores Ops Manager's own state. Sizing: [source]
- Small (< 50 MongoDB processes): 3-node RS, M10 equivalent (8 GB RAM) [source]
- Medium (50-500 processes): 3-node RS, M30 equivalent (32 GB RAM) [source]
- Large (500+ processes): 5-node RS or sharded cluster [source]
- App DB HA: Always use 3-node replica set. Single-node App DB = single point of failure for Ops Manager. [source]
MongoDB Agent
- A single binary that handles automation, monitoring, and backup for all MongoDB processes on the host. [source]
Automation (Declarative Goal State)
Backup Daemon
- Backup Daemon runs on a dedicated host. It: [source]
- Reads from oplog of source MongoDB (via agent) [source]
- Writes snapshots to configured store [source]
- Maintains PITR window by tailing the oplog [source]
- Place the Backup Daemon close to the data (low latency to both source MongoDB and snapshot store). [source]
Immutable S3 Snapshots (Object Lock)
- Enable S3 Object Lock on the S3 bucket to prevent snapshot deletion: [source]
Air-Gap Deployments (Local Mode)
Kubernetes Operator Integration
- Ops Manager provides a Kubernetes Operator (MongoDB Kubernetes Community Operator) for managing MongoDB deployments inside Kubernetes: [source]
Live Migration to Atlas
- Ops Manager supports initiating a Live Migration to Atlas: [source]
- Link Ops Manager to Atlas: Atlas UI → Live Migrate → Link to Ops Manager [source]
- Select source cluster: Choose the Ops Manager project and cluster [source]
- Configure Atlas target: Atlas project, cluster tier, region [source]
- Start migration: Ops Manager agent pulls data into Atlas [source]
- Cutover: Same mongosync-based cutover process [source]
Datadog
- Ops Manager pushes MongoDB metrics to Datadog for unified observability. [source]
PagerDuty
- Configure in Ops Manager UI → Alerts → PagerDuty integration. Maps Ops Manager alert categories to PagerDuty incident severity. [source]
Splunk
- Export Ops Manager logs via syslog or file-based log forwarding. Configure Splunk Universal Forwarder on Ops Manager hosts. [source]
Ops Manager Upgrade Path
- Backup App DB before upgrade [source]
- Check compatibility matrix: Ops Manager version → MongoDB Agent version → MongoDB server version [source]
- Upgrade Ops Manager application (rolling upgrade supported for multi-node OM deployments) [source]
- Upgrade MongoDB Agents on all hosts (Ops Manager prompts for agent upgrade) [source]
- Verify: Check agent connectivity status in Ops Manager UI [source]
- Version support policy: Ops Manager N, N-1, N-2 are supported. MongoDB Agent must be ≥ Ops Manager version. [source]
Anti-Patterns
- Single-node App DB: Ops Manager becomes unavailable if App DB node fails; always use 3-node RS [source]
- Backup Daemon on the MongoDB host: Backup creates I/O; place on dedicated host [source]
- Air-gap without pre-downloading all required MongoDB binaries: Automation will fail if the Agent can't find the requested version locally [source]
- Not using immutable S3 snapshots for compliance: Object Lock prevents accidental or malicious deletion of backup data [source]
- Manual edits to MongoDB configs outside Ops Manager automation: Ops Manager will reconcile these back to the goal state on next agent heartbeat [source]
References
Children
- Ops Manager Application (frontier)
- Application Database (App DB) (frontier)
- Backup Daemon and Head DB (frontier)
- MongoDB Agent (Automation, Monitoring, Backup) (frontier)
- Goal-State Automation Configuration (frontier)
- Continuous Backup with PITR (frontier)
- Blockstore Snapshot Storage (frontier)
- S3-Compatible Snapshot Storage with Object Lock (frontier)
- Filesystem Snapshot Storage (frontier)
- Local Mode and Air-Gap Deployments (frontier)
- Version Manifest Mirroring (frontier)
- Ops Manager LDAP/Kerberos/SAML/OIDC Federation (frontier)
- Workforce and Workload Identity Federation (frontier)
- Kubernetes Operator Deployment (frontier)
- Cloud Manager Free/Standard/Premium Tiers (frontier)
- Live Migration to Atlas via mongosync (frontier)
- Migration Host Provisioning (frontier)
- Source Oplog Window Sizing for Migration (frontier)
- PagerDuty/Datadog/Splunk Integration (frontier)
- OpenTelemetry MongoDB Receiver (frontier)
- Enterprise Advanced Licensing (frontier)
- Ops Manager High Availability (frontier)
- Cross-DC App DB Placement (frontier)
- Backup Daemon HA and Failure Domains (frontier)
Frontier under this node: Application Database (App DB), Backup Daemon HA and Failure Domains, Backup Daemon and Head DB, Blockstore Snapshot Storage, Cloud Manager Free/Standard/Premium Tiers, Continuous Backup with PITR, Cross-DC App DB Placement, Enterprise Advanced Licensing, Filesystem Snapshot Storage, Goal-State Automation Configuration, Live Migration to Atlas via mongosync, Local Mode and Air-Gap Deployments, Migration Host Provisioning, MongoDB Agent (Automation, Monitoring, Backup), OpenTelemetry MongoDB Receiver, Ops Manager Application, Ops Manager High Availability, Ops Manager LDAP/Kerberos/SAML/OIDC Federation, PagerDuty/Datadog/Splunk Integration, S3-Compatible Snapshot Storage with Object Lock, Source Oplog Window Sizing for Migration, Version Manifest Mirroring, Workforce and Workload Identity Federation