What an endpoint is

An endpoint stores the information that StorageGRID needs to use an external resource as a target for a platform service.

For example, to replicate objects from a StorageGRID bucket to an S3 bucket, the system needs to supply AWS with a bucket URI and credentials. This information is stored in an endpoint.

Each type of platform service requires its own endpoint, so you must configure at least one endpoint for each platform service you plan to use. After defining an endpoint, you use the endpoint's URN as the destination in the configuration XML used to enable the service.

You can use the same endpoint as the destination for more than one source bucket. For example, you could configure several source buckets to send object metadata to the same search integration endpoint so that you can perform searches across multiple buckets. You can also configure a source bucket to use more than one endpoint as a target, which enables you to do things like send notifications about object creation to one SNS topic and notifications about object deletion to a second SNS topic.

Endpoints for CloudMirror replication

StorageGRID supports replication endpoints that represent S3 buckets. These buckets might be hosted on Amazon Web Services, the same or a remote StorageGRID deployment, or another service.

To use an S3 bucket hosted on a StorageGRID system as an endpoint, specify an API Gateway Node in the endpoint definition.

Endpoints for notifications

StorageGRID supports Simple Notification Service (SNS) endpoints. Simple Queue Service (SQS) or AWS Lambda endpoints are not supported.

Endpoints for the search integration service

StorageGRID supports search integration endpoints that represent Elasticsearch clusters. These Elasticsearch clusters can be hosted in an AWS cloud or in a local datacenter.

The search integration endpoint refers to a specific Elasticsearch index and type. You must create the index in Elasticsearch before creating the endpoint in StorageGRID, or endpoint creation will fail. You do not need to create the type before creating the endpoint. StorageGRID will create the type if required when it sends object metadata to the endpoint.