Connection Strings Overview

• A connection string is a named, reusable object that stores the details RavenDB needs to connect to an external system or another RavenDB database, such as destination URLs, the target database or resource, and authentication parameters.

• RavenDB tasks and features reference a connection string by name.
  This keeps connection details outside each task or feature definition and lets the same connection string be reused.

• Connection strings can be defined in one of two scopes:

  • Per-database - available only to one database.
  • Server-wide - defined once at the cluster level and propagated automatically to databases in the cluster.

---

• In this article:
  • What connection strings are used for
  • Supported connection string types
  • Scope: per-database and server-wide
  • Managing connection strings

What connection strings are used for

Connection strings define the connection details that RavenDB tasks and features use to connect to another system:

• ETL tasks:
  Send transformed data from RavenDB to another RavenDB database, a relational database, cloud storage, a search engine, or a queue provider.
• External Replication:
  Replicate documents to another RavenDB database.
• Queue Sink tasks:
  Consume messages from a message broker and store the resulting documents in RavenDB.
• CDC Sink tasks:
  Stream changes from an external CDC database and store the resulting documents in RavenDB.
• AI features:
  Connect to AI providers used by embeddings generation, GenAI tasks, and AI Agents.

Supported connection string types

RavenDB supports the following connection string types:

Type	Connects to	Used by
RavenDB	Another RavenDB database	RavenDB ETL, External Replication, and Hub/Sink Replication
SQL	A relational database	SQL ETL and CDC Sink
Snowflake	A Snowflake warehouse	Snowflake ETL
OLAP	Local or cloud storage	OLAP ETL
Elasticsearch	An Elasticsearch cluster	Elasticsearch ETL
Queue	A supported queue broker or queue service	Queue ETL and Queue Sink tasks
AI	AI providers or RavenDB's embedded model	Embeddings generation, GenAI tasks, and AI Agents

Scope: per-database and server-wide

A connection string can be defined at one of two scopes:

Scope	Use when	Availability
Per-database	One database needs its own connection details, or different databases need different credentials, endpoints, or providers.	Stored in one database record and available only to tasks in that database.
Server-wide	Many databases share the same connection details, and you want one place to update endpoints, credentials, or provider settings.	Stored at the cluster level and propagated to all databases, except databases that are explicitly excluded.

Managing connection strings

Connection strings can be created, updated, read, and removed from Studio or through the Client API.

• Per-database connection strings
  Manage them from the database's Settings > Connection Strings view in Studio, or via the Client API.
  Learn more in:

  • Add or update a per-database connection string
  • Get per-database connection strings
  • Remove a per-database connection string

• Server-wide connection strings
  Manage them from Manage Server > Server-Wide Connection Strings in Studio, or via the Client API.
  Learn more in:

  • Add or update a server-wide connection string
  • Get server-wide connection strings
  • Remove a server-wide connection string

A connection string cannot be deleted while it is still referenced by a task.