Create an Atlas Vector Search Index with Kubernetes
On this page
Atlas Kubernetes Operator supports Atlas Vector Search indexes on Atlas clusters. To learn more, see How to Index Fields for Vector Search.
Prerequisites
Running Kubernetes cluster with Atlas Kubernetes Operator deployed.
MongoDB version
6.0.11,7.0.2, or higher on any cluster tier.Note
You can't create more than:
3 indexes on
M0clusters.5 indexes on
M2clusters.10 indexes on
M5clusters.
There are no limits to the number of indexes you can create on
M10+clusters.Collection to create the Atlas Vector Search index for.
Required Access
You need the Project Data Access Admin or higher role to
create and manage Atlas Vector Search indexes.
Procedure
Configure the AtlasDeployment Custom Resource.
Configure the spec.deploymentSpec.searchIndexes
object.
The following parameters apply to this object:
Parameter | Description |
|---|---|
DBName | Human-readable label that identifies the existing database that contains the collection with one or more Atlas Search indexes. ImportantThe database that you specify must already exist. If you don't have an existing database, create one. |
CollectionName | Human-readable label that identifies the collection that
contains one or more Atlas Search indexes. |
Type | Type of the index. Atlas accepts the following values:
|
To learn more about the available parameters, see
AtlasDeployment Custom Resource.
Example:
Run the following command.
cat <<EOF | kubectl apply -f - apiVersion: atlas.mongodb.com/v1 kind: AtlasDeployment metadata: name: my-atlas-cluster namespace: mongodb-atlas-system spec: projectRef: name: my-project deploymentSpec: backupEnabled: true clusterType: REPLICASET name: Test-cluster-M10 replicationSpecs: - regionConfigs: - backingProviderName: AWS electableSpecs: instanceSize: M10 nodeCount: 3 priority: 7 providerName: AWS regionName: US_EAST_1 zoneName: Zone 1 searchIndexes: - DBName: sample_airbnb collectionName: listingsAndReviews name: my-index vectorSearch: fields: | [ {"my-field":{"type": "vector", { ...}, ] type: vectorSearch terminationProtectionEnabled: false EOF