Managed MongoDB outside Atlas: in OSStend's cloud, your cloud or your data center
Fully managed MongoDB: OSStend engineers run replica sets and sharded clusters, backups and point-in-time recovery, monitoring, upgrades and security hardening, in OSStend's cloud, your cloud account or your data center.
45 minutes with an engineer: we review your workload and sketch the design.
- ✓Operated by AceMQ: 160+ customers, 10+ years
- ✓24/7 on-call engineers
An example design file, with illustrative values:
# Managed MongoDB service: mongo-catalog engine: mongodb topology: replica set, 3 members target: aws/byoc tier: fully managed backups: daily + oplog, pitr monitoring: included on_call: 24/7
MongoDB hosting from OSStend is a managed service run by on-call engineers: replica sets and sharded clusters with high availability, backups and point-in-time recovery, monitoring, upgrades and patching, security hardening and migration help. It runs in OSStend's cloud, your cloud account or your data center, and suits teams moving off Atlas or DocumentDB or running MongoDB themselves.
- PostgreSQL
- MySQL
- MongoDB
- Redis
- Valkey
- OpenSearch
- Kafka
- RabbitMQ
- AWS
- Google Cloud
- Azure
- Kubernetes
- OpenShift
- Prometheus
- Grafana
- Datadog
- Dynatrace
- Splunk
What does managed MongoDB from OSStend include?
The managed service covers the whole operating job for MongoDB: high availability, backups and point-in-time recovery, monitoring, upgrades and patching, security hardening, migration and 24/7 on-call. The table lists each part, with what MongoDB's own manual says about it.
| Included | What OSStend engineers do | What MongoDB documents |
|---|---|---|
| High availability | Design and run replica sets across zones or hosts, and sharded clusters where one replica set cannot hold the data or the write rate. Handle elections and failovers. | A replica set keeps the same data on several mongod processes and elects a new primary when one fails (checked 17 Nov 2026) |
| Replica sets and sharding | Size members, choose the shard key, run mongos routers and config servers, and rebalance when the data grows. | Every shard is a replica set, clients connect through mongos, and the shard key is hard to change (checked 17 Nov 2026) |
| Backups and point-in-time recovery | Daily backups plus the oplog, with recovery points set to your targets, and restore tests against them. | Filesystem snapshots and mongodump do not give continuous point-in-time recovery; Cloud Manager and Ops Manager read the oplog to do so (checked 17 Nov 2026) |
| Monitoring and on-call | Metrics and alerts forwarded to your Datadog, Dynatrace, Splunk or Prometheus. On-call engineers watch replication lag, the oplog window and cache pressure, around the clock. | n/a |
| Upgrades and patching | Version upgrades one major version at a time, minor releases and operating system patches, on your change calendar and never forced. | Major versions are upgraded one at a time; the security checklist says to track CVEs and end-of-life dates and apply OS patches (checked 17 Nov 2026) |
| Security hardening | Authentication and role-based access with least privilege, TLS for all connections, network exposure limited to trusted clients, disk or volume encryption where the design needs it, and a dedicated operating system user. | The pre-production security checklist lists each of these (checked 17 Nov 2026) |
| Migration help | OSStend runs the move off Atlas, DocumentDB or a self-run deployment and owns the cutover, with both sides live until you switch. | Atlas to self-managed MongoDB guide |
| Where it runs | In your cloud account or VPC, or in your data center on Kubernetes, OpenShift or bare metal, including air-gapped sites. | n/a |
The service commitments, uptime for fully managed and response time for co-managed, are set per plan and written into your contract. See service levels. Security measures are applied to the design you sign off; no attestation is claimed on this page.
What OSStend tailors for MongoDB
A MongoDB deployment is designed around the data and the access pattern: replica set or sharded cluster, sizing for data and indexes, the version, and where it runs. The design is signed off before anything is built, and the infrastructure stays yours.
- Backups and point-in-time recovery
- Metrics to your Prometheus or Datadog
- AceMQ engineers on call 24/7
| Area | Set for your workload |
|---|---|
| Size and shape | Instance size, storage and member count sized to your data and indexes, starting from how much of the index set and the hot data has to sit in memory. |
| Topology | A replica set elects a new primary when the old one is unreachable for more than 10 seconds by default, and the median election should not exceed 12 seconds, during which writes wait. Sharded clusters add mongos routers and a config server replica set, and every shard is itself a replica set. A cluster is sharded only when one replica set cannot hold the data or the write rate, and the shard key is chosen with that in mind. Zones and hybrid links as the workload needs. |
| Configs | The MongoDB and operating system settings the workload needs, including the default write concern and read concern, agreed with the team that owns the application. |
| Versions and upgrades | Major versions are upgraded one at a time: secondaries first, then a step-down and the old primary, with featureCompatibilityVersion raised only after a burn-in period. Version upgrades are handled by OSStend engineers and never forced. |
| Backups | Daily backups plus the oplog, with recovery points set to your targets. MongoDB describes mongodump and mongorestore as tools for small deployments with no continuous point-in-time restore, so larger sets use filesystem snapshots or a backup tool that can restore to a moment, and a sharded cluster adds coordination across shards. |
| Where it runs | Inside your cloud account or VPC, in your data center on Kubernetes or OpenShift, or at an air-gapped site. |
| Monitoring | Metrics and alerts forwarded to your Datadog, Dynatrace, Splunk or Prometheus. On-call engineers watch replication lag, the oplog window and cache pressure. |
There is no fixed menu here: the settings and cluster size your workload needs are written into the design you sign off before anything is built, and that design is what OSStend runs.
The MongoDB version clock, from the MongoDB support policy (Enterprise Advanced table, checked 9 Nov 2026): 6.0 ended on 31 July 2025, 7.0 ends on 31 August 2027 and 8.0 ends on 31 October 2029. A deployment still on 6.0 or earlier is outside that table, and reaching 7.0 or 8.0 means upgrading one major version at a time.
From our on-call engineers: the MongoDB settings that bite in production
MongoDB mostly goes wrong in storage, memory limits and replication lag, wherever it runs: OSStend's cloud, your cloud account or your data center. These are documented production notes we check on every replica set, and the first things to read after an unexplained slowdown.
- ✓Use XFS for WiredTiger data volumes, because MongoDB's production notes call it strongly recommended and warn that EXT4 can cause performance issues; disable atime on the volume.
- ✓The Transparent Huge Pages guidance depends on version: disable it on MongoDB 7.0 and earlier, enable it on 8.0.
- ✓The WiredTiger cache defaults to the larger of 50% of (RAM minus 1 GB) or 256 MB, and inside containers with less RAM than the host you set cacheSizeGB or cacheSizePct below the container limit.
- ✓The oplog defaults to 5% of free disk (between 990 MB and 50 GB), and a secondary that falls outside the oplog window cannot catch up by replication, so we check rs.printReplicationInfo() and can resize with replSetResizeOplog without a restart.
- ✓Low open-file limits cause connection failures under load, MongoDB warns at startup below 64,000, and clocks are kept in sync with chrony or NTP.
- ✓After one Postgres, Redis and MongoDB outage review, the common finding was that recovery had never been rehearsed, so we rehearse restores on MongoDB too.
MongoDB cloud hosting: in OSStend's cloud, your cloud account or your data center
MongoDB cloud hosting from OSStend means engineers run MongoDB on dedicated, single-tenant clusters in OSStend's cloud, inside your AWS, Azure or Google Cloud account, or in your data center. Choose your own account or data center when the data, the billing and the security boundary have to stay with you.
| Option | What it is | Fits when |
|---|---|---|
| MongoDB Atlas | MongoDB's own managed service. Clusters deploy onto AWS, Google Cloud or Azure and are charged per hour by cluster tier, for each data-bearing node (checked 17 Nov 2026). | The data may live in a vendor-run service and a tier fits the workload. Prices and limits are on the MongoDB Atlas pricing and alternatives page. |
| Amazon DocumentDB | A separate AWS service that implements the MongoDB API on its own engine (checked 9 Nov 2026). | The workload stays inside AWS and works with the documented API differences. |
| MongoDB on cloud VMs, self-run | You run mongod on your own instances, and backups, upgrades and on-call are your team's. | You have a DBA team and want full control. |
| OSStend, in OSStend's cloud | Dedicated, single-tenant MongoDB clusters that OSStend engineers run for you: replica sets or sharded clusters, backups and point-in-time recovery, monitoring, upgrades and 24/7 on-call. | You want MongoDB run end to end without owning the infrastructure, and a standard Atlas plan does not fit. |
| OSStend, in your cloud account | Engineers operate MongoDB inside your account and VPC: replica sets or sharded clusters, backups and point-in-time recovery, monitoring, upgrades and 24/7 on-call. AWS comes first, then Azure and Google Cloud. | The data and the bill must stay in your account, and you want an on-call team to run it. |
| OSStend, in your data center | The same service on hosts you own, on Kubernetes, OpenShift or bare metal, with air-gapped sites included. | The workload must stay on-premises or is moving back from the cloud. |
How we compared: MongoDB's and AWS's own documentation, dated in the table. Atlas and DocumentDB are the better choice when a plan already fits and costs less than running the nodes yourself. OSStend is our service.
Where turnkey MongoDB services stop
Plan-based MongoDB services restrict what the operator can do. Atlas blocks commands it automates, and Amazon DocumentDB implements the MongoDB API on its own storage engine instead of running MongoDB. Teams outgrow them on cost, on control or on where the data has to live.
| Area | What a plan fixes | What OSStend designs instead |
|---|---|---|
| Engine | DocumentDB emulates the MongoDB 3.6, 4.0, 5.0 and 8.0 APIs on a purpose-built engine, so query plans and explain() output can differ and there is no admin or local database (checked 9 Nov 2026). | MongoDB itself, with member count and storage sized from your data and indexes. |
| Commands | Atlas blocks commands such as eval, applyOps, mapReduce and replSetReconfig because it automates what they do (checked 9 Nov 2026). | The settings and replica set changes your workload depends on, agreed in the design and made by OSStend engineers. |
| Upgrade timing | The provider's schedule. | Your change calendar, one major version at a time. |
| Where it runs | The provider's infrastructure. | Your cloud account or your data center, with the infrastructure under your control. |
Plan menus differ by provider and change over time. The MongoDB Atlas alternative page compares costs and limits with sources and dates.
Change data capture from MongoDB
Debezium has a MongoDB source connector, so changes in a replica set or sharded cluster can stream to Kafka or other sinks. OSStend runs the connector and watches the oplog window against connector downtime.
See managed Debezium for how the pipeline is run, and change data capture explained for the methods and tools.
Design. Move. Run. for MongoDB
OSStend works in three steps for MongoDB. An engineer designs the deployment and you sign it off, OSStend synchronises the new members and owns the cutover, and named engineers then operate it 24/7 on your infrastructure.
Design
An engineer reviews your workload (data size, indexes, access patterns, recovery targets and where it must run) and writes the design: replica set or sharded layout, sizing, shard key, version, backups and security settings. You sign it off before anything is built.
Move
OSStend runs the migration off Atlas, DocumentDB or a self-run deployment and owns the cutover. The new deployment is synchronised from the source, with the method set in the design step, and both sides stay live until you switch. A DocumentDB source is checked against its API differences first, and applications are tested against MongoDB before traffic moves. Zero downtime is the plan.
Run
Named engineers run it 24/7: backups and restore tests, version upgrades and patching on your calendar, index and cache tuning, elections and failovers, security upkeep and incident response, with response-time targets written into your contract.
Deployment and service levels
MongoDB is delivered in OSStend's cloud, in your cloud account and VPC, or in your data center. OSStend engineers operate the database in all three.
In OSStend's cloud
OSStend operates MongoDB on dedicated, single-tenant clusters in its own cloud, with the same engineers and the same service as in your environment.
Fully managedIn your cloud account
OSStend operates MongoDB inside your account and VPC, so the data, the billing and the security boundary stay with you. AWS first, then Azure and Google Cloud.
Fully managedIn your data center
Hosts you own, on Kubernetes, OpenShift or bare metal, with air-gapped sites included. Delivered through an operator on your cluster, with supported charts and images.
Co-managed
Fully managed deployments carry an uptime commitment. Co-managed deployments carry a 24/7 response-time commitment, because you own the infrastructure. The service levels page sets out the tiers, and deployment shows each option in detail.
Uptime and response figures are set per plan and written into your contract. AceMQ, which operates OSStend, sets its support response targets from 15 minutes, by plan.
MongoDB Community Server is published under the Server Side Public License (checked 31 Oct 2026). OSStend runs MongoDB in its own cloud, in your cloud account or in your data center.
Proof from AceMQ's operating record
The elections, backups and version upgrades for MongoDB are run by AceMQ's engineers. AceMQ operates OSStend, has 160+ customers and 10+ years of running open-source infrastructure, and keeps engineers on call 24/7, as acemq.com states.
Customers stay unnamed unless they have given AceMQ written permission, so the MongoDB record here is anonymous. A comparable replica set or sharded estate can be discussed in the 45-minute review.
Pricing for managed MongoDB
Managed MongoDB is priced from four drivers: the deployment design, its size, the SLA tier and where it runs. An engineer prices the design after a workload review, and there is no free tier or card sign-up for MongoDB.
- Deployment design
- a three-member replica set or a sharded cluster with config servers and routers asks for a different number of members.
- Cluster size
- member count, instance size and storage, which follow data and index size.
- SLA tier
- fully managed in OSStend's cloud or your cloud account, or co-managed in your data center, and the commitment attached to each.
- Where it runs
- OSStend's cloud, your cloud account or your data center. In your own account or data center, the infrastructure bill stays with you.
Say which MongoDB version you run, how large the data and indexes are and whether it is sharded, and the engineer prices that deployment.
The pricing page shows how the model applies across every engine.
When to contact OSStend, and when not to
Contact OSStend when you want an on-call team to operate MongoDB in OSStend's cloud, your account or your data center. If Atlas or DocumentDB still does the job at a price you accept, stay there.
Contact us if
- ✓Your Atlas or DocumentDB bill, limits or location no longer fit, and you want MongoDB run for you, in OSStend's cloud or in an account or data center you control.
- ✓MongoDB must run in your cloud account or in your data center, including air-gapped sites, or you want it on dedicated clusters in OSStend's cloud.
- ✓You run MongoDB yourself and want a 24/7 on-call team, upgrades and backups handled.
- ✓You want the migration off Atlas run for you.
Not a fit if
- ✕An Atlas or DocumentDB plan already fits your workload, and it costs less than running the nodes yourself.
- ✕You want a free tier, or to sign up with a card. For MongoDB, OSStend starts with a conversation with an engineer.
- ✕You want a free tier or self-serve sign-up. OSStend does not offer either for MongoDB.
Questions about managed MongoDB hosting
These are the questions engineering and finance teams ask most before moving MongoDB off Atlas or handing it to a managed team.
What is MongoDB hosting?
MongoDB hosting supplies the servers that run MongoDB. Managed MongoDB hosting adds the operations: replica sets and sharding, backups and point-in-time recovery, monitoring, upgrades and patching, security hardening and 24/7 on-call engineers. OSStend provides the managed kind, in OSStend's cloud, your cloud account or your data center.
What is MongoDB cloud hosting, and does it need Atlas?
No. MongoDB cloud hosting means MongoDB running on cloud infrastructure, and Atlas is one way to get it. OSStend engineers run MongoDB on dedicated clusters in OSStend's cloud, or inside your own AWS, Azure or Google Cloud account so the data and the bill stay with you. The comparison with Atlas shows the trade-offs.
What does the managed MongoDB service cover?
High availability, replica sets and sharded clusters, backups and point-in-time recovery, monitoring, upgrades and patching, security hardening, migration help and 24/7 on-call. The table near the top of this page lists each part. Application schema and queries stay with your team.
Can you run sharded MongoDB clusters?
Yes. A sharded cluster has shards that are each a replica set, mongos routers and a config server replica set, and OSStend designs the shard key and runs all of it. A cluster is sharded only when one replica set cannot hold the data or the write rate.
How do backups and point-in-time recovery work?
Daily backups plus the oplog, with recovery points set to your targets and restores tested against them. MongoDB's manual lists continuous point-in-time recovery for oplog-based tools such as Cloud Manager and Ops Manager, and none for filesystem snapshots or mongodump (checked 17 Nov 2026), so the tool is chosen in the design.
What happens when the primary fails?
The replica set holds an election, which the secondaries start once they lose contact with the primary for more than 10 seconds by default. MongoDB says the median election should not exceed 12 seconds with default settings, and writes wait until it completes. Applications need a retry policy and timeouts that outlast the election, and the design checks that.
How does a move off Atlas or DocumentDB work?
The new deployment is synchronised from the source while both sides stay live, with the method set in the design step. Traffic switches only after the new side is verified, and OSStend engineers own the cutover. The MongoDB Atlas alternative page describes the path, and the Atlas to self-managed MongoDB migration guide covers the phases and the version limits.
Which MongoDB versions can I run?
The version is set in the deployment design and agreed before anything is built. MongoDB's lifecycle table lists 7.0 to 31 August 2027 and 8.0 to 31 October 2029 (checked 9 Nov 2026), and upgrades go one major version at a time on your change calendar.
How does the cost compare with Atlas?
The price follows member count, instance size, the SLA tier and where it runs, and the infrastructure bill stays with you. The MongoDB Atlas alternative page compares the models, and an engineer prices the specific design after a workload review.
Should this workload be on MongoDB or PostgreSQL?
That depends on whether the data is naturally documents or relational. The MongoDB vs PostgreSQL comparison sets out the documented differences, and the design review checks the choice against your own data model.
Related pages
- Migration guideMigrate MongoDB off Atlas without downtimeMigrate MongoDB off Atlas with mongosync or mongodump: version limits, what to check, how cutover works, and managed running by AceMQ engineers.
- Service levelsService levels: two tiers, a written SLAOSStend service levels: fully managed and co-managed tiers, what each one covers, who is on call and how incidents are handled, with a written SLA for each.
- AlternativeMongoDB Atlas pricing and alternatives in 2026MongoDB Atlas pricing and alternatives: Atlas list prices, and managed MongoDB in OSStend's cloud, your cloud or your data center, with the migration run for you.
- Change data captureManaged Debezium: change data capture run 24/7Managed change data capture with Debezium on Kafka Connect or Debezium Server, run 24/7 in OSStend's cloud, your cloud account or your data center.
- GuideChange data capture (CDC): methods and toolsChange data capture explained: log-based vs query-based CDC, the main CDC tools, Postgres CDC and replication slots, and when to pay someone to run it.
- DeploymentManaged databases: our cloud, yours or on-premManaged databases in OSStend's cloud, your AWS, Azure or Google Cloud account, or your data center. PostgreSQL, Kafka and RabbitMQ, run 24/7 by engineers.
Talk to an engineer about your MongoDB workload
45 minutes with an engineer: bring your MongoDB version, the size of your data and indexes and where the cluster runs now, and the engineer sketches the design.
- ✓A design sketch for your workload, not a plan menu
- ✓A zero-downtime cutover path for what you run today
- ✓A price for that design, after the review