# Atlas Kubernetes Operator AKO v2.14 > Source-anchored research for Atlas Kubernetes Operator AKO v2.14. Run status: COMPLETED. Parent: [MongoDB Atlas Infrastructure as Code](https://llms-explorer.com/tree/mongodb-atlas-infrastructure-as-code/) · 3 facets · 20 facts · page: https://llms-explorer.com/tree/atlas-kubernetes-operator-ako-v2-14/ ## How it works - The v2.14.0 release notes report fixes for an infinite AtlasDeployment reconciliation loop when terminationProtectionEnabled was enabled through the UI, incorrect MongoDB connection-URI parsing for connection secrets, and reconciliation getting stuck on ebsVolumeType. These are maintainer-reported fixes, not independent runtime validation. (low confidence) — [source](https://github.com/mongodb/mongodb-atlas-kubernetes/releases/tag/v2.14.0) - AKO dry run previews create/update/delete actions as Kubernetes events without applying those Atlas changes. The documented upgrade-testing procedure uses a new Kubernetes cluster because only one AKO CRD version can exist in a cluster. (medium confidence) — [source](https://www.mongodb.com/docs/atlas/operator/current/ak8so-dry-run.md) ## Comparisons and alternatives - The October changelog snapshot lists v2.15, v2.16, and v2.17 releases after v2.14; v2.14 is a historical release target rather than the newest version represented by that documentation. (low confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/operator/current/ak8so-changelog.md) ## Facts and statements - AKO manages Atlas cloud database resources through Kubernetes APIs; it does not run the Atlas database inside the Kubernetes cluster. (medium confidence) — [source](https://www.percona.com/blog/mongodb-operators-explained-features-limitations-and-open-source-alternatives/) - AKO 2.14.0 introduces custom resources generated from the MongoDB Atlas Administration API OpenAPI specification. Its tagged reference uses atlas.generated.mongodb.com/v1 and lists Cluster, DatabaseUser, FlexCluster, Group, and IPAccessListEntry. (high confidence) — [source](https://github.com/mongodb/mongodb-atlas-kubernetes/releases/tag/v2.14.0) - The v2.14.0 generated Cluster specification supports v20250312. It allows at most one versioned specification, nests create/update fields under entry, and makes groupId and groupRef mutually exclusive. (low confidence) — [source](https://github.com/mongodb/mongodb-atlas-kubernetes/blob/v2.14.0/docs/api-docs-generated.md) - Current documentation says generated and existing CRDs are incompatible for managing the same Atlas resource. It describes generated coverage as limited and expanding, with existing CRDs intended for eventual replacement; this is current guidance, not a dated v2.14 deprecation deadline. (low confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/operator/current/generated-crds-overview.md) - The 2.14.1 patch fixes incorrect annotations on FlexCluster and DatabaseUser that prevented their export. This is a documented patch-level distinction from 2.14.0. (low confidence) — [source](https://github.com/mongodb/mongodb-atlas-kubernetes/releases/tag/v2.14.1) - The v2.14.0 and v2.14.1 release notes both list Kubernetes 1.33–1.35 and OpenShift 4.21 as supported versions for those releases; those historical release notes do not establish today's support policy. (high confidence) — [source](https://github.com/mongodb/mongodb-atlas-kubernetes/releases/tag/v2.14.0) - The documented v1-to-v2 upgrade changes YAML field names and passes credentials through Kubernetes Secrets. It moves management to a new Kubernetes cluster and scales the old operator to zero before the replacement takes over. (low confidence) — [source](https://www.mongodb.com/docs/atlas/operator/current/upgrade-ako-v1-to-v2.md) - In issue #2869, a user running AKO 2.11.0 reported that two AtlasIPAccessList CRs for one project overwrite each other's entries. Maintainers explained that the CR owns the whole list and recommended one list CR per project. The thread does not test v2.14. (low confidence) — [source](https://github.com/mongodb/mongodb-atlas-kubernetes/issues/2869) - A Huma engineer's 2021 issue reported unwanted project deletion and removal of manually entered access-list IPs while adopting existing projects, later identifying v0.5.1. Another participant could reproduce only the IP-list overwrite. These are historical, partially disputed reports, not confirmed v2.14 defects. (low confidence) — [source](https://github.com/mongodb/mongodb-atlas-kubernetes/issues/265) - Percona's May 2, 2025 comparison describes AKO as Apache-2.0 operator code controlling Atlas services and distinguishes operator licensing from the usage charges of the underlying Atlas service. The article is a competitor's comparison and does not assess v2.14. (low confidence) — [source](https://www.percona.com/blog/mongodb-operators-explained-features-limitations-and-open-source-alternatives/) - AKO applies the Kubernetes operator pattern to Atlas infrastructure: custom resources describe desired configuration, and a controller reconciles it with Atlas. Kubernetes documents that CRDs alone store structured data; reconciliation requires a controller. (medium confidence) — [source](https://www.mongodb.com/docs/atlas/operator/current/generated-crds-overview.md) - MongoDB explicitly prohibits using generated and existing CRDs to manage the same Atlas resource; choose one CRD set per managed resource. (low confidence) — [source](https://www.mongodb.com/docs/atlas/operator/current/generated-crds-overview.md) - The current generated-CRD documentation lists Group, Cluster, DatabaseUser, FlexCluster, and IpaccessListEntry. This is an October documentation snapshot, not a verified inventory of the v2.14.0 binary. (low confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/operator/current/generated-crds-overview.md) - AKO dry run predates v2.14: the changelog introduced it in v2.8.0. It previews proposed Atlas create/update/delete requests as Kubernetes events without applying those Atlas changes. (medium confidence) — [source](https://www.mongodb.com/docs/atlas/operator/current/ak8so-changelog.md) - For upgrade dry runs, MongoDB instructs users to use a new Kubernetes or OpenShift cluster, install the target-version CRDs, and apply resources plus credentials there. Its rationale is that replacing CRDs in the existing cluster can break reconciliation. (low confidence) — [source](https://www.mongodb.com/docs/atlas/operator/current/ak8so-dry-run.md) - The dry-run documentation marks the feature public preview and illustrates an operator image tagged 2.8.0. Treat that example as historical configuration, not a v2.14 image pin. (low confidence) (verified as of 2026-10-14) — [source](https://www.mongodb.com/docs/atlas/operator/current/ak8so-dry-run.md) - Percona’s May 2025 comparison emphasizes that AKO controls the external Atlas service and does not run database servers inside the Kubernetes cluster. Its cost and lock-in criticism is a competing vendor’s perspective, not evidence of a v2.14 runtime defect. (low confidence) — [source](https://www.percona.com/blog/mongodb-operators-explained-features-limitations-and-open-source-alternatives/)