Flex Limitations and migration to Dedicated
Parent: Atlas Pricing Programs, Cost Governance and TCO · Published reference · snapshot 2026-10-21 · skill ranked-3d355de6
↓ Facts as markdownall context files
Source-anchored research for Flex Limitations and migration to Dedicated. Run status: COMPLETED.
These notes link each claim to its source. Some citations identify archived research reports that are no longer hosted on this site. A published reference means the content is available; it does not certify independent review or accuracy.Read the editorial policy and follow the sources before relying on a claim.
How it works
- Flex does not support private endpoints or VPC peering, so clients must use public endpoints. Atlas recommends Dedicated clusters when private connectivity is required, and a MongoDB Community response confirms that Flex clients connect through the public endpoint. (high confidence) (verified as of 2026-10-14) [source]
- Atlas does not support scaling a Dedicated cluster back to Free or Flex, so treat the tier conversion as one-way when selecting the target. (low confidence) (verified as of 2026-10-14) [source]
- Two public user reports describe Flex-to-Dedicated upgrades failing during migration or rollback: one Flex-to-M20 report in August 2025 and one Flex-to-M10 report in September 2026. The latter author says they created a new M10 and migrated manually. These reports establish that failures have been reported, but they do not establish cause or frequency. (medium confidence) [source]
Problems, failure modes and limitations
- Atlas Flex limits total data storage to 5 GB of uncompressed BSON plus indexes, caps read and write throughput at 500 operations per second, and allows at most 500 connections. Above 500 operations per second, Atlas applies a hard limit and cooldown or queuing; it does not offer a higher Flex price tier to buy more throughput. Treat either the storage or throughput threshold as a Dedicated migration trigger. (high confidence) (verified as of 2026-10-18) [source]
- Flex does not support Continuous Backup or point-in-time restore, sharded clusters, or dedicated workload-isolation nodes. Flex snapshots are daily and cannot use custom schedules or on-demand snapshots; Dedicated clusters can add configurable backups and point-in-time restore, and dedicated nodes require M10 or higher. (high confidence) (verified as of 2026-10-14) [source]
- Flex clusters cannot be paused while idle. Dedicated clusters can be paused for up to 30 days, but storage remains billable and backups stop while paused. Scheduled idle windows can therefore change the TCO comparison. (low confidence) (verified as of 2026-09-30) [source]
- Atlas's native tier-change path is one-way: an M10+ Dedicated cluster cannot be scaled down to Flex or Free. Scaling Flex to M10+ requires an initial sync and downtime; Atlas says to stop writes for the duration, and notes that migration can affect performance while nodes resync. (high confidence) (verified as of 2026-10-20) [source]
- Do not treat an Atlas tier upgrade as Atlas Live Migration (Pull): MongoDB says Flex clusters do not support that feature and cannot be its destination. For a separate source-to-destination move into Atlas, AWS guidance describes deploying the destination, live migrating data, and cutting over; that is a different path from scaling the existing Flex cluster. (high confidence) (verified as of 2026-10-21) [source]
Comparisons and alternatives
- Atlas's current Flex pricing guidance lists $8 to $30 per 30 days, billed hourly, with 5 GB included; Dedicated cost varies with tier, provisioned storage, and region. Flex's price cap also bounds capacity: paying more does not remove the 500-operations-per-second ceiling. Compare the Dedicated tier and storage needed for the workload rather than treating the Flex cap as elastic capacity. (high confidence) (verified as of 2026-10-18) [source]
- For cost-governance triage, PointFive recommends comparing peak rather than average operations-per-second and connection metrics to Flex limits, checking storage and feature requirements, and retaining Dedicated where limits or Dedicated-only features matter. It identifies small development, staging, CI, demo, and proof-of-concept databases as Flex candidates only when they fit those constraints. (medium confidence) (verified as of 2026-10-21) [source]
Facts and statements
- Atlas supports in-place scaling from Flex to Dedicated through the Administration API's flexClusters:tenantUpgrade endpoint. Atlas Kubernetes Operator also supports in-place migration with upgradeToDedicated; its procedure requires billing to be enabled, the full target Dedicated configuration, and the same cluster name. (high confidence) (verified as of 2026-10-14) [source]
- The documented Flex-to-Dedicated scale operation requires downtime, and Atlas instructs operators to halt writes during the change to prevent data corruption. A MongoDB Community response also describes a short outage for one upgrade, so schedule maintenance and check the estimate shown for the specific cluster. (high confidence) (verified as of 2026-10-14) [source]
- For a separate-cluster Atlas migration, AWS Prescriptive Guidance describes deploying a new target cluster, live-migrating the data, and cutting over. This is general Atlas migration guidance, not a Flex-specific tier-conversion procedure. (low confidence) [source]
- Flex lacks capabilities that can require Dedicated, including private endpoints and VPC peering, continuous backup and point-in-time restore, sharding, customer-managed encryption keys, configurable storage or memory, and dedicated search or analytics nodes. A workload that depends on one of these features needs a Dedicated tier even if its current storage and throughput are low. (high confidence) (verified as of 2026-10-14) [source]
- For an Atlas Kubernetes Operator-managed Flex cluster, MongoDB documents an in-place migration by setting `spec.upgradeToDedicated: true` and providing the full target `spec.deploymentSpec`. The cluster name must remain the same, project billing must be enabled, and the operator reports progress through the `DeploymentReady` condition. (high confidence) (verified as of 2026-10-14) [source]
- The MongoDB Atlas Terraform Provider v2.10 migration guide documents changing the tier through the Atlas UI, then importing the Dedicated cluster as `mongodbatlas_advanced_cluster`. It recommends backing up Terraform state and removing the old `mongodbatlas_flex_cluster` state entry so Terraform does not plan to delete the existing cluster. (low confidence) (verified as of 2026-10-21) [source]
Children
- Flex backup and recovery planning (frontier)
- Terraform state reconciliation after a Flex-to-Dedicated conversion (frontier)
Frontier under this node: Flex backup and recovery planning, Terraform state reconciliation after a Flex-to-Dedicated conversion