Docs Menu
Docs Home
/
MongoDB Atlas
/ /

Query using Pre-Defined Replica Set Tags

On this page

  • Pre-Defined Replica Set Tag Descriptions
  • Disk States
  • Node Types
  • Use Cases and Examples
  • Use Analytics Nodes to Isolate Workloads
  • Target Local Reads for Geographically-Distributed Applications
  • Reduce Secondary Disk Warming Impact
  • Retrieve Availability Zones
  • Built-In Custom Write Concerns

Note

This feature is not available for M0 free clusters, M2, and M5 clusters. To learn more about which features are unavailable, see Atlas M0 (Free Cluster), M2, and M5 Limits.

Atlas clusters are configured with pre-defined replica set tags for different member types in the cluster. You can utilize these pre-defined replica set tags to direct queries from specific applications to specific node types, regions, and availability zones. These pre-defined replica set tags allow you to customize read preferences for a replica set, thus improving cluster performance and reliability.

Note

These pre-defined replica set tags differ from the resource tags that you provide and manage. You can't change these replica set tags that Atlas provides.

To use pre-defined replica set tags in your connection string and direct queries to specific nodes, use the following connection string options:

  • readPreference

  • readPreferenceTags

  • readConcernLevel

For examples, see Use Cases and Examples.

The following table describes the pre-defined replica set tags that Atlas provides.

Pre-Defined Tag Name
Description
Example
Availability Zone

AWS availability zone ID, Google Cloud fully-qualified name for a zone, or Azure zone number.

Azure supports availability zones only in a subset of regions. Atlas provides pre-defined availability zone tags for Azure only for regions that support availability zones. To learn more, see Microsoft Azure.

For more information about the possible availabilityZone values for each cloud provider, see the cloud provider's documentation:

  • AWS: {"availabilityZone" : "use1-az1"}

  • Google Cloud: {"availabilityZone" : "us-east1-b"}

  • Azure: {"availabilityZone" : "1"}

Node Type

Node type.

Possible values are:

  • ELECTABLE

  • READ_ONLY

  • ANALYTICS

For more information, see Node Types.

{"nodeType" : "ANALYTICS"}
Provider

Cloud provider on which the node is provisioned.

Possible values are:

  • AWS

  • GCP

  • AZURE

{"provider" : "AWS"}
Region

Cloud region in which the node resides.

For a complete list of possible region values, see the reference page for your cloud provider:

{"region" : "US_EAST_2"}
Workload Type

Pre-defined replica set tag to distribute your workload evenly among your non-analytics (electable or read-only) nodes.

Possible values are:

  • OPERATIONAL

{"workloadType" : "OPERATIONAL"}
Disk State

State of your disk.

Possible value: READY

For more information, see Disk States.

{"diskState" : "READY"}

The following table describes the possible diskState values in your pre-defined replica set tags.

Disk State
Description
READY

Only target warm nodes.

You can run queries without experiencing increased or unpredictable latency.

For an example of this replica set tag, see Reduce Secondary Disk Warming Impact.

The following table describes the possible nodeType values in your pre-defined replica set tags.

Node Type
Description
ELECTABLE
Read from nodes eligible to be elected primary. ELECTABLE nodes correspond to Electable nodes for high availability in the cluster creation UI.
READ_ONLY
Read from read-only nodes. READ_ONLY nodes correspond to Read-only nodes for optimal local reads in the cluster creation UI.
ANALYTICS
Read from read-only analytics nodes. ANALYTICS nodes correspond to Analytics nodes for workload isolation in the cluster creation UI.

To learn how to configure electable, read-only, and analytics nodes for your cluster, see Configure High Availability and Workload Isolation.

Tip

See also:

For details on how these pre-defined replica set tags correspond to BI Connector for Atlas read preferences, refer to the BI Connector cluster options section of the Create a Cluster Page.

Consider the following scenarios where utilizing pre-defined replica set tags would be beneficial, and see the corresponding sample connection strings.

Note

Each of the following example connection strings employ the readConcernLevel=local connection string option. Specifying a read concern of local ensures that secondary reads on sharded clusters do not return orphaned documents. You do not need to specify this option when connecting to non-sharded replica sets.

If an application performs complex or long-running operations, such as ETL or reporting, you may want to isolate the application's queries from the rest of your operational workload by connecting exclusively to analytics nodes.

Consider the following connection string:

mongodb+srv://<db_username>:<db_password>@foo-q8x1v.mycluster.com/test?readPreference=secondary&readPreferenceTags=nodeType:ANALYTICS&readConcernLevel=local

The connection string options appear in the following order:

  • readPreference=secondary

  • readPreferenceTags=nodeType:ANALYTICS

  • readConcernLevel=local

The readPreference option of secondary and readPreferenceTag option of { nodeType : ANALYTICS } limit the application connections to analytic nodes.

You may want to isolate regular application reads from the workload on analytics nodes.

Consider the following connection string:

mongodb+srv://<db_username>:<db_password>@foo-q8x1v.mycluster.com/test?readPreference=secondary&readPreferenceTags=workloadType:OPERATIONAL&readConcernLevel=local

The connection string options appear in the following order:

  • readPreference=secondary

  • readPreferenceTags=workloadType:OPERATIONAL

  • readConcernLevel=local

The specified options prevent your application from reading from analytics nodes. The application must read from operational, or non-analytics, nodes.

Use pre-defined replica set tags to target local reads to specific regions for globally distributed applications. Prior to the introduction of these pre-defined replica set tags, local reads for globally distributed applications relied on correctly calculating the nearest read preference. With pre-defined replica set tags, specifying appropriate geographic tags in combination with a read preference mode of nearest provides more consistent behavior.

The following connection string prioritizes connections to the AWS US_EAST_1 region, followed by the US_EAST_2 region:

mongodb+srv://<db_username>:<db_password>@foo-q8x1v.mycluster.com/test?readPreference=nearest&readPreferenceTags=provider:AWS,region:US_EAST_1&readPreferenceTags=provider:AWS,region:US_EAST_2&readPreferenceTags=&readConcernLevel=local

The connection string options appear in the following order:

  • readPreference=nearest

  • readPreferenceTags=provider:AWS,region:US_EAST_1

  • readPreferenceTags=provider:AWS,region:US_EAST_2

  • readPreferenceTags=

  • readConcernLevel=local

Atlas considers each read preference tag in the order you specify them. After Atlas matches a node to a tag, it finds all eligible nodes that match that tag. Atlas then ignores any following readPreferenceTags.

In this example, the application first tries to connect to a node in AWS region US_EAST_1. If there no nodes in that region are available, the application tries to connect to a node in AWS region US_EAST_2.

The final (empty) readPreferenceTags= provides a fallback option. With an empty readPreferenceTags= option, the application can connect to any available node regardless of provider or region.

These options help ensure that the application connects to the closest geographic region for reduced latency and improved performance.

Note

You can further target reads based on availability zones.

Tip

See also:

For additional information and use cases for various read preferences, refer to the Read Preference page in the MongoDB Manual.

When Atlas adds or replaces a node in your cluster, it performs disk pre-warming by default. During the disk pre-warming process, the newly created storage volume undergoes a period of heavy IOPS usage. This slows down read operations made against this node. Therefore, during the disk pre-warming process, Atlas keeps the pre-warming node hidden by default, preventing it from participating in any read operations.

If you prefer that a newly added or replaced node becomes immediately active and visible, you can choose to disable fast disk pre-warming and prevent read operations on a warming node by using a connection string as in the following example:

mongodb+srv://<db_username>:<db_password>@foo-q8x1v.mycluster.com/test?readPreference=secondary&readPreferenceTags=diskState:READY&readConcernLevel=local

The connection string options appear in the following order:

  • readPreference=secondary

  • readPreferenceTags=diskState:READY

  • readConcernLevel=local

The readPreference option of secondary and readPreferenceTag option of { diskState : READY } tells Atlas to only target warm nodes.

Atlas provides pre-defined replica set tags for availability zones by default. These tags include the AWS availability zone ID, Google Cloud fully-qualified name for a zone, or Azure zone number. You can check the availability zone of a node by using the rs.conf() shell method or by viewing the last ping of the cluster node.

Example

...
"tags": {
"availabilityZone": "use2-az2",
...

Atlas provides built-in custom write concerns for multi-region clusters. Write concern describes the level of acknowledgment requested from MongoDB for write operations to a cluster.

Atlas's built-in custom write concerns can help improve data consistency by ensuring your operations are propagated to a set number of regions to succeed.

To use a custom write concern, specify the write concern in the write concern document of your operation.

Atlas provides the following custom write concerns for multi-region clusters:

Write Concern
Tags
Description
twoRegions
{ region: 2 }
Write operations must be acknowledged by at least two regions in your cluster.
threeRegions
{ region: 3 }
Write operations must be acknowledged by at least three regions in your cluster.
twoProviders
{ provider: 2 }
Write operations must be acknowledged by at least two regions in your cluster with distinct cloud providers.

Example

Consider a multi-region cluster across three regions: us-east-1, us-east-2, and us-west-1. You want to have write operations propagate to all three regions in your cluster before Atlas accepts them.

The following operation inserts a document and requires that the operation be propagated to all three regions due to the { w: "threeRegions" } write concern object:

db.employees.insertOne(
{ name: "Bob Smith", company: "MongoDB" },
{ writeConcern: { w: "threeRegions" } }
)

Back

HA and Workload Isolation